U.S. patent application number 14/011172 was filed with the patent office on 2013-12-26 for client processor device.
The applicant listed for this patent is Freestyle Technology Pty Ltd. Invention is credited to David Lewis Beard, Andrew Paul Donaghey, Ian Kenneth McDonald.
Application Number | 20130346570 14/011172 |
Document ID | / |
Family ID | 52574860 |
Filed Date | 2013-12-26 |
United States Patent
Application |
20130346570 |
Kind Code |
A1 |
Donaghey; Andrew Paul ; et
al. |
December 26, 2013 |
Client Processor Device
Abstract
A client processor including a communications receiver for
receiving file fragments for the device, a micro-controller system
including a microcontroller and a programmable non-volatile memory
system, for building and storing application and data files from
the fragments, and executing an application of the device by
processing at least one application file and associated data
identified by configuration instructions included in at least one
of the fragments.
Inventors: |
Donaghey; Andrew Paul;
(Vermont South, AU) ; McDonald; Ian Kenneth;
(Victoria, AU) ; Beard; David Lewis; (Huntingdale
Victoria, AU) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Freestyle Technology Pty Ltd |
Mitcham Victoria |
|
AU |
|
|
Family ID: |
52574860 |
Appl. No.: |
14/011172 |
Filed: |
August 27, 2013 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
13275978 |
Oct 18, 2011 |
|
|
|
14011172 |
|
|
|
|
11571198 |
Dec 22, 2006 |
8041792 |
|
|
PCT/AU05/00922 |
Jun 24, 2005 |
|
|
|
13275978 |
|
|
|
|
60608379 |
Sep 10, 2004 |
|
|
|
Current U.S.
Class: |
709/221 |
Current CPC
Class: |
G06F 9/44563 20130101;
G06F 16/1847 20190101; H04L 41/0886 20130101 |
Class at
Publication: |
709/221 |
International
Class: |
H04L 12/24 20060101
H04L012/24 |
Foreign Application Data
Date |
Code |
Application Number |
Jun 24, 2004 |
AU |
2004903457 |
Sep 13, 2004 |
AU |
2004905262 |
Claims
1. A client processor device, including: a communications receiver
for receiving fragments of files for said device, without
requesting the file fragments; a microcontroller system for
processing said file fragments and receiving and storing multiple
files at once, and automatically performing an application defined
by said files without any user intervention or interaction.
2. A client processor device as claimed in claim 1, wherein a new
application, as defined by said files, is performed by the device
without any user intervention or interaction.
3. A client processor device as claimed in claim 1, wherein the
device is reconfigured, as defined by said files, without any user
intervention or interaction.
4. A client processor device as claimed in claim 1, wherein the
communications receiver is a wireless communications receiver.
5. A client processor device as claimed in claim 4, wherein a new
application, as defined by said files, is performed by the device
without any user intervention or interaction.
6. A client processor device as claimed in claim 4, wherein the
device is reconfigured, as defined by said files, without any user
intervention or interaction.
7. A client processor device as claimed in any one of the preceding
claims, wherein the received fragments for the device include an
initial file defining the remaining files as being of a type of
application, configuration or data, and defining reading of the
files.
8. A client processor device as claimed in claim 7, wherein an
operating system of the microcontroller system includes at least
one of: (i) a clock task for implementing a real time clock and
calendar; (ii) a virtual machine task for initiating a virtual
machine, and causing reading of said initial file; and executing of
a file; (iii) a command task for establishing device commands for
production testing and initial configuration; (iv) a file system
task for processing said fragments and said files and removing
unwanted and redundant files; (v) a communications interface task
for controlling said receiver and processing data received by the
receiver; and (vi) tasks for processing of input and output signals
for I/O peripherals, such as input buttons, or a GPS receiver of
said client processor device.
9. A client processor device as claimed in claim 8, wherein the
operating system includes a scheduler to allocate each task to be
processed by the microcontroller system, and allocates semaphores
to resources of said device to allow the microcontroller system to
identify resources allocated to a running task.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a division of U.S. application Ser. No.
13/275,978, filed Oct. 18, 2011 (now allowed), which is a
continuation of U.S. National application Ser. No. 11/571,198, now
U.S. Pat. No. 8,041,792, having a 371(c) date of Dec. 22, 2006,
which is a 371 of International Application No. PCT/AU05/00922,
filed Jun. 24, 2005, which claims priority from U.S. Provisional
Application No. 60/608,379, filed Sep. 10, 2004 (now expired). The
contents of all of these applications are incorporated herein by
reference in their entirety.
TECHNICAL FIELD
[0002] The present invention relates to a client processor device
that is able to remotely receive application, configuration or data
files to establish, adjust or control the application executed by
the device.
BACKGROUND
[0003] Consumer electronic devices are now commonly used for a wide
variety of applications. Many of the devices now include an
embedded processor architecture, which normally comprises a
microprocessor and associated memory. Instructional code and data
for the device is stored in the architecture, and defines and
controls the application for which the device has been specifically
built. For example, a compact disc (CD) player and a cordless
telephone each include a basic processor architecture, but the
similarities then normally cease as the CD player is specifically
configured to perform a number of tasks associated with processing
the audio data on a CD, whereas a telephone is specifically
configured to perform a number of call handling tasks. Similar
considerations apply to other electronic devices, such as video
game consoles, mobile telephones, electronic toys, cameras, etc.
The device is specifically configured for its chosen application,
and that application cannot be changed, particularly by the receipt
of any remote instructions. Even more general processing devices,
such as a wireless personal digital assistant (PDA) or a personal
computer connected to a computer network, have defined operating
systems with a defined set of APIs that restrict the applications
that can be performed on the devices, and restrict the extent to
which the devices are remotely configurable. Restrictions are also
imposed by the security constraints associated with the network
that can be used to connect to the device, and any configuration is
performed on a one to one basis, ie only one client device can be
configured remotely at a given time.
[0004] It is desired to address the above or at least provide a
useful alternative.
SUMMARY
[0005] In accordance with the present invention there is provided a
client processor device including:
[0006] a communications receiver for receiving file fragments for
said device;
[0007] a microcontroller system including a microcontroller and a
programmable non-volatile memory system, for building and storing
application and data files from said fragments, and executing an
application of said device by processing at least one application
file and associated data identified by configuration instructions
included in at least one of said fragments.
[0008] The present invention also provides a client processor
device, including:
[0009] a wireless communications receiver for receiving files for
said device;
[0010] a microcontroller system for processing and storing said
files, and automatically performing an application defined by said
files without any user intervention or interaction.
[0011] The present invention also provides a client processor
device, including:
[0012] a microprocessor;
[0013] a programmable non-volatile memory system addressable by the
microprocessor;
[0014] a communications receiver for receiving file fragments for
said device; and
[0015] a microcontroller system for processing a first application
based on stored application data of said device and subsequently
processing a second application based on said transmitted
application data.
DESCRIPTION OF DRAWINGS
[0016] Preferred embodiments of the present invention are
hereinafter described, by way of example only, with reference to
the accompanying drawings, wherein:
[0017] FIG. 1 is a block diagram of a preferred embodiment of a
client processor device;
[0018] FIG. 2 is a block diagram of compilation components for
firmware of the client processor device;
[0019] FIG. 3 is a block diagram of compilation components for an
application of the client processor device;
[0020] FIG. 4 is a block diagram of firmware management components
for the client processor device;
[0021] FIG. 5 is a block diagram of the logic architecture of the
client processor device;
[0022] FIG. 6 is a block diagram of the software architecture of
the client processor device;
[0023] FIG. 7 is a block diagram of a flash memory system of the
client processor device;
[0024] FIG. 8 is a block diagram of the flash file system
architecture of the client processor device;
[0025] FIG. 9 is a block diagram of an application management
system of the client processor device;
[0026] FIG. 10 is a block diagram of a file transmission system of
the client processor device;
[0027] FIG. 11 is a block diagram of an erasure correction fragment
generation process of the client processor device;
[0028] FIG. 12 is a block diagram of a Lagrange polynomial
interpolation process of the client processor device;
[0029] FIG. 13 is a flow diagram of an initialisation procedure for
the flash memory system;
[0030] FIG. 14 is a flow diagram of a directory entry
initialisation procedure for the initialisation procedure of the
flash memory system;
[0031] FIG. 15 is a flow diagram of a receive process of the flash
memory system;
[0032] FIG. 16 is a flow diagram of an update quanta process of the
receive process of the flash memory system;
[0033] FIG. 17 is a block diagram of the client device receiving
files from the server transmission system via a variety of network
interfaces;
[0034] FIG. 18 is a flow diagram of a file house keeping process of
the flash memory system;
[0035] FIG. 19 is a flow diagram of an incomplete file removal
process of the flash memory system;
[0036] FIG. 20 is a flow diagram of a file deletion process of the
flash memory system;
[0037] FIG. 21 is a flow diagram of a building missing fragments
process of the flash memory system;
[0038] FIG. 22 is a block diagram of a remote relay unit for
communicating with the client processor device;
[0039] FIG. 23 is a block diagram of a server transmission system
for communicating with the client processor device;
[0040] FIG. 24 is a block diagram of the client processor device
for a call management application;
[0041] FIG. 25 is a block diagram of a logical architecture of the
device for the call management application;
[0042] FIG. 26 is a block diagram of a carrier selection main
process loop of the system for the call management application;
[0043] FIG. 27 is a block diagram of an initialisation process of
the system;
[0044] FIG. 28 is a flow diagram of an exchange code resolution
process of the system for the call management application;
[0045] FIG. 29 is a flow diagram of a file loader process of the
system for the call management application;
[0046] FIG. 30 is a flow diagram of a best national carrier process
of the system for the call management application;
[0047] FIG. 31 is a flow diagram of a redirection option process of
the system for the call management application; and
[0048] FIG. 32 is a flow diagram of an interface option process of
the system for the call management application.
DETAILED DESCRIPTION
1. The Client Processor Device
[0049] A client processor device 100, as shown in FIG. 1, has a
housing 101 that contains a printed circuit board on which three
semiconductor chips 110, 120 and 150 are mounted to provide a
receiver 110, a non-volatile programmable memory 120, in the form
of data-flash memory, and a microcontroller 150. Other support
components for the chips are included on the board within the
housing 101, such as a battery power system, and any additional
input/output (I/O) components 130 that are required, such as a LED
or LCD display and control input buttons or keys. The
microcontroller 150 is connected to the receiver 110 and the memory
120 via a serial interface 117 and to an address and data bus 114
for connection to the I/O components 130, and has an asynchronous
serial interface 115 for external peripherals. The housing 101 is
relatively small, and the footprint of the device 100 occupies less
than 20 cm.sup.2. The electronic components of the device 100 also
consume less than 100 mW.
[0050] The microcontroller 150 is an 8-bit microcontroller with
flash read only memory (ROM) 106 (about 128 kbytes),
electronically, erasable and programmable read only memory (EEPROM)
109 and random access memory (RAM) 108 (less than 4 kbytes), such
as the PIC 18 Series Microcontrollers produced by Microchip
Technology Inc (http://www.microchip.com). The processor core 102
of the microcontroller 150 has a hardware return stack of 32
levels, and no formal data stack, making it primarily suitable to
handling only one processor task. Offset addressing is not
supported and only two hardware interrupts are provided, one low
and one high priority interrupt, without any support for software
interrupts. The return stack stores a return address only.
[0051] The memory 120 provides data storage on a data flash device,
such as the AT45 Series of data-flash memories produced by Atmel
Inc (http://www.atmel.com), using a hierarchal structure of
sectors, blocks and pages as discussed in detail below. A complete
page, comprising 264 or more bytes depending on device type, is
written at once, whereas at least one block is erased, with a block
comprising eight pages of data. The data-flash memory 120 is
managed by a flash file system, described in detail below.
[0052] The receiver 110 is an RF transceiver, such as the NT29xx
series produced by Numa Technologies, Inc
(http://www.numatechnologies.com), and is controlled by the
microcontroller 150 (via the serial interface 117) to provide an
analog output 118, for sampling by the processor core 102, of the
radio frequency (RF) signal received by the receiver 110. The
receiver provides an air interface, and is frequency agile being
phase locked loop synthesised. The receiver uses diversity antennas
111 and 112 connected to the transceiver chip 110, under control of
the processor 150 using a control signal 113 to switch the signal
from one or the other antenna, to improve performance under
marginal signal conditions. A switching arrangement is used, but it
is also possible to use dual receivers and receive paths, which may
be tuned to different frequencies.
[0053] The firmware production process is shown in FIG. 2. The
firmware 2003 stored in the ROM 106 of the device 100 is primarily
written as source code 2001 in the C programming language and
compiled using the PICC-18 ANSI C compiler 2002 provided by HI-TECH
Software LLC. The compiler uses a compiled stack to deal with the
omitted data stack from the processor core 102. The data stack
would normally hold auto-allocated (local) variables and function
parameters. Instead, the complete calling hierarchy of the code is
evaluated at compile time by the compiler to produce a call-graph
tree, as described below. The auto-allocated variables and function
parameters are allocated to a fixed memory bank in a section of the
RAM 108, being designated parameter RAM by the compiler, and the
allocations are shared where possible to minimise the overall
parameter RAM requirements. For example, the maximum possible
parameter RAM size may be 256 bytes, a compromise between the bank
switching limitations of the processor and the speed and size of
the code. This maximum allocation size is normally more than
sufficient for a single main task and routines called using the low
and high priority interrupt handlers. The compiled stack is only
designed to handle a single task without supporting re-entrant
code. The compiler defines three separate roots in a call graph,
one for main execution and one for each of the interrupts.
[0054] The application production process is shown in FIG. 3. An
application 2012 sent to and stored in the flash memory 106 of the
device 100 is primarily written as source code 2010 in a
programming language, such as C, and compiled into byte code for a
virtual machine (VM) using a compiler 2011 for the device 100.
Operating System (OS)
[0055] The operating system (OS) 210 can be considered to reside on
and be supported by the hardware 202 of the device 100, as shown in
FIG. 5. The code for the OS and its basic tasks 220 are stored in
the flash ROM 106. The OS on boot, sets the interrupts, initialises
semaphores (described below) and scheduling, and supports and runs
a start-up task 227 which in turn starts the remaining basic tasks
220, including the virtual machine task 221 on which the
application code 230 relies. The number of tasks is limited to
eight due to the word length of one byte of the processor 102 and
because an increase would substantially increase overhead. The OS
provides a cooperative multi-tasking environment to allow
asynchronous events to be handled independently, with much reduced
code complexity, enabling the device 100 to be more responsive and
reliable. The asynchronous events are normally those either caused
externally to, or not under control of, the device, such as button
presses, characters received on a serial port interface of the I/O
130 or a hardware peripheral connected to the I/O 130 completing an
operation. Most of the events require a response in a timely
manner.
[0056] The operating system is able to share the time of the
processor 102 between the tasks by rapidly context switching
between them on a cooperative basis. The tasks are: [0057] (i) A
clock task 225. This maintains time of day and a calendar. [0058]
(ii) A virtual machine (VM) task 221. This processes the byte code
of the application files as discussed below. [0059] (iii) A command
task 222. This establishes a set of commands available via the
local asynchronous serial interface 115. [0062] (iv) A flash file
system task 223. This handles building and management of all files
within the flash memory 120. [0060] (v) An air task 224. This
establishes communications and control of the transceiver 110 and
supports an air protocol of the device 100, as discussed below.
[0061] (vi) Additional tasks 226. This represents any other tasks
as required, for example, to process data received from a global
positioning system (GPS) receiver. [0062] (vii) The start-up task
227. This establishes the remaining tasks in a controlled fashion
as discussed below.
[0063] The tasks are each assigned a unique priority depending upon
the required processing rate and latency, when started. When a
context switch is performed, the OS moves the processor stack
contents to a memory area allocated for the task being swapped out,
and then restores it with the stored stack for the new task to be
run. By allocating stack memory levels required for each task,
memory requirements are contained. The storage level can be
allocated from 2 to 32 depending on the task, when it is started. A
context switch may involve calling a function (e.g. the OS_Yield( )
function) in the OS, which causes the running task to yield control
to the OS and therefore allowing another task to run. Preferably,
context switching is achieved only by calling the OS_Yield( )
function. A common scratch-pad area of the RAM 108 is used for
temporary processing and is only relevant within a function call,
as the contents are ignored if another routine is called, which may
overwrite the contents. Accordingly, when the OS accomplishes a
task switch, the scratch pad area does not need to be saved thereby
providing a significant saving of RAM and processor overhead.
[0064] Many software modules and hardware resources (e.g. the
data-flash memory 120, the EEPROM 109, and the bus 117) need to be
shared between tasks, and each shared resource is allocated a
binary semaphore (or token). If a resource is used and another task
attempts to use it, then this is prohibited until the first task
finishes with the resource. Locking of the shared resources
protects them from re-entrant calls. In the device 100, normally 20
to 30 semaphores are allocated to lock the resources. Binary
semaphores are controlled as a bit field in bit addressable memory
of the RAM 108, and counting semaphores are used to control linear
or multi-level resources (such as buffers). The semaphores are
allocated when the firmware 2003 is compiled by the compiler
2002.
[0065] To protect the shared code (and resource) via locking, the
following procedure is used: [0066] (a) Request access to the
module/resource via the semaphore. [0067] (b) Lock out other tasks
using the semaphore associated with the module/resource once access
is obtained; [0068] (c) Call functions for the module/resource; and
[0069] (d) Allow other tasks access by freeing the semaphore
associated with the module/resource.
[0070] These steps (a) and (d) may be implemented as functions in
the corresponding software modules to save code space (e.g. as Lock
and Unlock functions).
[0071] Task re-entrant code can be written to allow multiple tasks
to call the same routine. A particular task may only call the one
routine once. To do this the routine's parameters are saved if they
must be accessed after the context switch. This can be achieved
using an array to provide context sensitive storage for each task
as follows:
TABLE-US-00001 char storage[8]; // Parameter storage for each of 8
tasks ... void shared(char var) { storage[<current_task>] =
var; // save the parameter ... OS_Yield( ); // context switch var =
storage[<current_task>] ; // Restore parameter... }
[0072] The OS also includes a scheduler that is responsible for
determining the processor time allocated to each task. The highest
priority task (e.g. task "1") is allocated up to 50% of the
processor time, given the opportunity to run every second context
switch, whilst the next high office priority task (e.g. task "2")
is allowed to run 25% percent of the context switches and so on.
For example, the task may be switched based on a priority level as
follows: 1 2 1 3 1 2 1 4 1 2 1 3 1 2 1 5, etc. If a task cannot be
run because it is blocked, then the next lowest priority task is
run and so on. This allows resource intensive or critical tasks to
be run more often or with lower latency whilst ensuring slower
tasks are not able to stall. The scheduler in addition is able to
prevent a potential priority inversion hang without requiring a
complex priority inheritance procedure. While lower priority tasks
are guaranteed to run, a higher priority task waiting on a resource
in use by a lower level task will not be indefinitely blocked.
[0073] The low and high priority interrupts of the processor 100
are treated as being slow and fast, respectively. The fast
interrupt is reserved for handling time critical tasks, such as
analog to digital conversion of a sampled analog signal. The fast
interrupts generally do not relate to the use of shared routines or
those of the OS, and the slow interrupts are used to handle
asynchronous events, such as timer interrupts, serial port events
and lower level packet decoding that directly relate to OS
controlled tasks. The low interrupt handlers use OS functions, such
as semaphores to communicate with tasks.
[0074] A timer module 316 of the OS provides, shown in FIG. 6,
accurate timing of intervals and delays, and allows tasks to be
synchronised to a period such as one second for a clock of the
device 100. The timer module allows a variety of different high and
low resolution durations to be measured, and a number of timers,
can be allocated when the firmware is compiled.
Software Architecture
[0075] The software architecture employed by the firmware stored in
the ROM 106 distinguishes between software modules 300 and the
tasks 220, as shown in FIG. 6. The software modules 300 are
available for use by the tasks 220, the arrows indicating
dependencies. An object-oriented approach is used in the module
decomposition process, and some modules are given responsibility at
a given time for a particular shared resource (e.g. the serial
interface, EEPROM, etc.) as required.
[0076] There are four classes of software modules 300. The first
type 310 includes the OS core 312 that manages scheduling and task
switching, and the semaphore 314 and timing 316 modules. These
modules are typically used by all tasks and written to guarantee
that they can be called at any time (i.e. not require locking). The
second type are shared modules that must be locked to ensure
conflicts do not occur. These include an SPI module 323 that
manages data transfer on the serial peripheral interface and an EE
module 320 that allows access to the processor's EEPROM 109. These
modules typically contain a number of available routines, which are
locked as a group (i.e. these routines are locked/unlocked using
the same semaphore).
[0077] The third type are modules run autonomously from a task.
These include the virtual machine 321, command processor 322, air
interface 324 and start-up 327 modules. These are each dedicated to
a particular task, not shared and therefore do not require
locking.
[0078] The fourth type is a combination of the latter two. These
have components that are run only from a single task, for example
in the flash file system (FFS) module 330, the module 331 is run
from a FFS task 223 whereas a set of shared functions, 332 and 333
are available for use by external modules, including the internal
component 331. The shared functions are locked for use. The clock
module 340 is another example, where a dedicated module 341 is run
from the clock task 225, and a set of shared functions 342, 343 are
also available for use by other modules.
[0079] The main( ) routine 328 is responsible for initializing all
modules as required and calling an OS 312 routine to run the
start-up task 227
Call Graph Management
[0080] Manipulation of the call graph supports a multi-tasking
environment. Call-graph allocation is performed using assembler
directives unique to the firmware compiler 2002. These include
directives to define a root, and to allocate each routine to a
particular root. Any of the modules that have contained routines
that run to completion and do not require a separate call-graph
root are defined to use a default main root to save parameter
space
[0081] A flattening call-graph approach could be used but typically
results in the very limited parameter space allocation being easily
exceeded. Rather, the call graph is broken into a number of
separate trees, whereby the main function of each task forms the
base of a tree or root. This allows the overlaying of auto
variables between functions and parameter passing. For routines
that are not run to completion, assembly directives in the task
code are used. These are used internally by the compiler to ensure
interrupt code starts at a separate base and interrupts can run at
any time. Routines or functions that are shared across tasks are
always assigned to a root. Shared functions which always run to
completion between context switches may all share the same root as
they will not run concurrently, but they may not share a root with
tasks that do not always run to completion. Any shared functions
that may incur a context switch before they complete are assigned a
unique root. Groups of such functions or routines that are
guaranteed never to be called concurrently by more than one task
can be assigned the same root to conserve parameter space.
[0082] Any functions solely called either directly or indirectly by
the main( ) routine 328 need not be formally placed in the
call-graph as they will be placed in the main root of the
call-graph by the compiler. These would typically include once-off
initialization and configuration functions.
Firmware Management
[0083] The device 100 can be structured so that only a small boot
routine 160 is permanently held in the device ROM 106 as depicted
in FIG. 4. The boot routine 160 can program the remaining firmware
components 170 into the ROM with data obtained via the local
interface 115, or transfer execution to the remaining firmware
components once they are programmed.
[0084] Once the remaining firmware components are in place and
operating, the flash file system of the client processor device 100
receives and stores files in the flash memory 120 for processing by
the virtual machine 321 of the OS. The files normally relate to an
application or a version of the application, but the files may also
include updates to the remaining firmware components. The boot
routine 160 may read file information from the flash memory 120 but
without the need to implement the flash file system to the full
extent. The boot routine may compare the versions of the firmware
components programmed in the device ROM with files identified to be
firmware (e.g 121, 122) in the flash memory 120 and automatically
update the firmware components in the device ROM, if it determines
they are newer.
[0085] The device 100 is entirely remotely configurable, without a
user having to interact directly with the device, e.g. by selecting
a button.
Flash File System
[0086] The flash file system of the client processor device 100
receives and stores files in the flash memory 120 for processing by
the virtual machine 321 of the OS. The flash file system 400, as
shown in FIG. 7, is a linear file system that allows up to 64 files
of up to 64 kbyte to be stored in the flash memory 120. The file
system allows storage of files (e.g. files 411, 412), in a
write-protected area 410 and optionally in a read-write area 420
(e.g. file 421, 422, 423, 424), as shown in FIG. 7. The
write-protected area is read-only if the device is not in an
initial configuration state, otherwise write access is allowed. The
flash file system supports 512 kbyte and 1 Mbyte data flash devices
120 and this does not preclude other device sizes or numbers of
devices, if required.
[0087] The data-flash memory 120 is divided up into 264 byte pages,
where 256 bytes are used for actual data, and the 8 spare bytes are
used for indicating what portions of the page contain valid data. A
structure is used, as shown in FIG. 8, where 508 represents RAM
components and 504 the components stored in the data-flash memory
120. One or two flash blocks are treated as an allocation unit,
depending on the device capacity, so that a total of 256 allocation
units are available. An allocation unit may be a directory unit
502, or a data unit 505. The directory units 502 each include
directory entries 510 having a header 512 and a table 509. The
header 512 includes information about the file, such as its local
creation date 541, CRC 542, release date 543, length 544, data
quanta count 545, erasure quanta count, 546, access mode 547, name
548, type 549, quanta size 550, quanta count 551, erasure
correction 552. Table 514 includes pointers 509 to the data
allocation units 505 that actually hold the data for the file. Each
data allocation unit holds one or two blocks, being eight or
sixteen pages 530 depending on the flash device capacity. A page
530, containing valid data, consists of 256 bytes of data 531, 4
bytes of quanta 532, and a 32-bit CRC. The quanta 532 represents up
to 32 indicator flags, which indicate which bytes of the data 531,
are valid. The RAM 108 maintains a set of tables 520, which each
have pointers 507 to predefined number of directory units 502 which
each include 16 or 32 directory entries 510 per unit, depending on
flash device capacity.
[0088] Allocation units in write-protected flash and those in
read-write flash are allocated from pools 521 and 522,
respectively, representing free allocation units, maintained in RAM
108.
[0089] A table 523 is maintained for the state of all directory
entries 510 in the directory allocation units 502 in the
write-protected and read-write areas of the flash memory 120. The
state defines whether the directory entry 510 contains an allocated
file header 512, and if so, the file's name and flags indicating,
whether it is complete, and any request that it be deleted. For a
complete file additional state flags indicate whether it is
released, for use by the virtual machine. The state also indicates
whether a released the file is locked and/or open.
[0090] A count of the total number of writes to each allocation
unit is maintained in a table in EEPROM 109. The count is used to
ensure the next allocation unit with the lowest count is used
thereby achieving even wear of the flash. If this count exceeds a
threshold, then the file system will not reallocate the allocation
unit, and it can be treated as allocated in order to avoid further
access to it. A free allocation unit with the least number of
writes to it is allocated first. The file system updates the
pointers accordingly if allocation units are moved due to a worn
unit.
[0091] The flash file system task 223 on initialisation executes a
start up procedure 7000, as shown in FIG. 13. The start up
procedure 7000 begins at step 7002 by searching the write-protected
area 410 of the flash memory 120 for the expected number of
directory allocation units 502. Step 7004 determines whether such
directory allocation units are found. If all units are found, step
7004 proceeds to step 7006 to search the read-write area 420 of the
flash memory 120 for directory allocation units. Then, if step 7008
determines that directory allocation units are found in the
read-write area, step 7008 proceeds to step 7080, where all
directory entries are processed.
[0092] However, if step 7004 determines that the expected number of
directory allocation units are not found in the write-protected
area, step 7004 proceeds to step 7024, to determine whether the
unit is in an initial configuration state. Initial configuration
occurs at manufacture being indicated by a dedicated hardware
signal. If step 7024 determines that the unit is not being
configured, step 7026 records the error and, at step 7028, a signal
is generated to indicate that the flash file system has failed to
start.
[0093] If step 7024 determines that the unit is being configured,
step 7024 proceeds to step 7030 to format the flash memory 120
(e.g. by deleting data from both the write-protected and read-write
areas of the memory 120). At step 7032, header units in the
write-protected area are created, and at step 7036, header units in
the read-write area are similarly created. Step 7036 proceeds to
step 7040 to process all directory entries, as shown in FIG.
14.
[0094] If step 7008 determines that directory allocation units 502
are not found in the read-write area, step 7008 proceeds to step
7034 to format only the read-write area. Step 7034 proceeds to step
7036 to create header units in the read-write area. Step 7036
proceeds to step 7040 to process all directory entries, as shown in
FIG. 14.
[0095] Once all the directory entries have been processed, step
7040 proceeds to step 7010 to determine whether any of the
directory allocation units 502 in the read-write area of the memory
120 are missing. If none are missing, step 7010 proceeds to step
7022 to start the file system. Otherwise, step 7010 proceeds to
step 7012 to search for a free allocation unit in table 52 (e.g. in
the read-write area). Step 7014 determines whether a free
allocation unit has been found. If found, step 7014 proceeds to
step 7016 to assign the free allocation unit as a directory entry.
Step 7016 the proceeds to step 7010 to continue processing other
directory entries. If step 7014 determines that a free allocation
unit cannot be found, step 7014 proceeds to step 7018 to record the
error, and at step 7020, generates a signal indicating that the
file system has failed to start.
[0096] The process shown in FIG. 14 processes all directory entries
in order to ascertain the files resident in the flash memory, the
state and integrity of each file (i.e. whether complete, incomplete
or corrupted) and the allocation units used. No assumptions are
made about the state or number of files, or data allocation units
prior to this process.
[0097] Step 7042 is the first step for individually processing each
directory entry 510 in the directory allocation units 502 in the
write-protected area, and also the directory allocation units in
the read-write area of memory 120. Each directory entry 510
potentially contains data related to a file (e.g. 411, 421). Step
7042 selects a directory entry for processing. At step 7044, the
file name 548 represented by the selected directory entry
(contained in header of the selected directory entry), and the file
location (e.g. write-protected or read-write), and attributes such
as completeness are entered into a RAM directory table 523. Step
7052 then determines whether the selected directory entry contains
file data. If step 7052 determines that the file contains no data,
step 7052 proceeds to step 7046 which determines whether all the
directory entries identified by the directory allocation units (eg.
502) in memory 120 have been processed. If so, step 7048 proceeds
to step 7050 indicating all directory entries have been processed.
Otherwise, step 7046 proceeds to step 7048, which selects another
directory entry from the directory allocation units in memory and
proceeds to step 7044.
[0098] If step 7052 determines that the file contains data, step
7052 proceeds to step 7054 to select a data allocation unit 505
corresponding to the file (e.g. by retrieving the pointer to the
first data allocation unit (e.g. 505) from the table (e.g. 514) for
the selected directory entry). Step 7054 then proceeds to step 7052
to process the selected data allocation unit by determining whether
that unit has been allocated in memory 120 by checking table 522.
If so, step 7056 proceeds to step 7058 to first determine whether
the file can entry can be deleted from the corresponding directory
allocation unit (as the allocation unit is incorrectly in use by
another file) by determining whether the file is in write-protected
space. Otherwise, step 7056 proceeds to step 7064 to check the
contents of the selected data allocation unit.
[0099] If step 7058 determines the file is in write-protected
space, step 7058 proceeds to step 7060 which determines if the
device is in an initial configuration state. Otherwise, step 7058
proceeds to step 7062 to clear the file entry. If step 7060
determines the device is not in an initial configuration state,
step 7060 proceeds to step 7094 to indicate an unrecoverable error.
Otherwise, step 7060 proceeds to step 7062 to clear the file entry
by de-allocating all data allocation units referred to in the by
the file's header table 514, except the data allocation unit
already in use by another file, then clearing the file's directory
entry 510. Step 7062, then proceeds to step 7046 which determines
whether all the directory entries have been processed. If so, step
7046 proceeds to step 7050 indicating that all directory entries
have been processed. Otherwise, step 7046 proceeds to step 7048 to
select a new directory entry and proceeding to step 7044.
[0100] Step 7064 proceeds to check the contents of the selected
allocation unit 505 by selecting the unit's first page and
proceeding to step 7066. Step 7066 determines whether the selected
page 530 contains any data by checking whether the data quanta
flags 532 indicate any data bytes are present. If so, step 7066
proceeds to step 7068 to check the CRC of the selected page by
evaluating the CRC of the data 531 and quanta flags (e.g. 532)
Otherwise, step 7066 proceeds to step 7078 to check whether all the
pages in the allocation unit have been processed.
[0101] Step 7068 proceeds to step 7070 to check whether the
evaluated CRC in step 7068, matches the stored CRC 533. If not,
step 7070 proceeds to step 7072 to record the error and, at then
proceeds to step 7058 to delete the file entry from the
corresponding directory allocation unit, if allowed. Otherwise,
step 7070 proceeds to step 7074 to set a flag in table 522
indicating that the data allocation unit 305 has been
allocated.
[0102] Step 7074 proceeds to step 7076 to record a running total of
the amount of data quanta contained in the data pages, and proceeds
to step 7078 to check whether all the pages in the allocation unit
have been processed. If not, step 7078 proceeds to stop 7080 to
select another page in the data allocation unit for processing, and
proceeds to step 7056. Otherwise, step 7078 proceeds to step 7082
to determine whether all the data allocation units for the file (as
defined in the directory entry) have been processed. If not, step
7082 proceeds to step 7094 to select another data allocation unit
for the same file (e.g. another pointer to a data allocation unit
for that file), and the new selected data allocation unit is
processed at step 7056.
[0103] If step 7082 determines that all the data allocation units
for the file have been processed, step 7082 proceeds to step 7084.
Although all the data allocation units for a file (as defined in
the directory entry) have been processed, the file may still be
incomplete. Step 7084 determines whether the data for the file is
complete by whether the quanta accumulated in step 7076 matches the
expected quanta count (e.g. 545) in the file header 540. If not,
step 7084 proceeds to step 7046 to select another directory entry.
Otherwise, step 7084 proceeds to step 7086 to check the CRC data
for the file as a whole. Step 7088 determines whether the CRC check
at step 7086 indicates that all the data contained in the file is
consistent with the file CRC 543 contained in the header 540. If
not, step 7088 proceeds to step 7058 to delete the file entry from
the corresponding directory allocation unit, if allowed. Otherwise,
step 7088 proceeds to step 7090.
[0104] Step 7090 determines whether the file presently being
processed is a duplicate of another file in the flash file system
by comparing file header (e,g, 512) parameters. If step 7090
determines that the file is not a duplicate, step 7090 proceeds to
step 7046 to select a new directory entry. Otherwise, step 7090
proceeds to step 7092 to determine whether the date/time stamp on
the present file 543 represents a time that is earlier than the
date/time of another file of which the present file is a duplicate.
If so, step 7092 proceeds to step 7058 to check if the file entry
can be deleted from the corresponding directory allocation unit.
Otherwise, step 7092 proceeds to step 7046 to 7046 to check whether
all the directory entries have been processed
[0105] The failures identified in flash memory initialisation
processes 7000 and 7040 may occur as a result of corruption or loss
of data in one or more flash memory pages. Possible causes of such
corruption may be power failure or system re-boot whilst the data
is being written to the device 100, or random data bit-errors due
to wear and/or cross-talk between pages. It is beneficial to guard
against these modes of failure and reducing their severity if they
do occur. This can be achieved by modifying the file system to:
[0106] (a) wear the flash memory as evenly as possible; [0107] (b)
ensure loss of a single data page never affects more than one file;
and [0108] (c) ensure corruption, or loss, of a single page can be
recovered.
[0109] The first two are satisfied by employing distributed file
directory headers throughout the flash memory (e.g. within the
first data allocation unit of each file), rather than aggregating
them in a few allocation units. The last can be addressed, at least
in part, by using duplicate directory headers for each file, each
on separate flash memory pages, to allow a corrupted header to be
recovered from the intact header. To recover a lost page from the
data area of a file, a simple erasure correction can be implemented
by storing an additional page (e.g. along with the duplicated
directory headers), to provide a page-wise parity of all data pages
in the file. A corrupt header or data page can be identified by the
failure of the CRC associated with each page written to flash.
[0110] The flash file system processes file fragments received by
the receiver 110 to perform a receive process 7100, as shown in
FIG. 15. Space in the flash 120 is only allocated for the actual
fragments received, to reduce overhead, and each file fragment is
time stamped with its local creation date 541, in addition to a
release date 543 determined by the transmitting server. Fragments
can be received in different orders and multiple files received at
once, which allows for interleaved transmission and prompt
transmission of high priority files.
[0111] The receive process 7100 as shown in FIG. 15 begins at step
7102 on receiving a file fragment. Step 7102 searches the RAM
directory table 523 for an unreleased file (i.e. an incomplete file
or completed file not in the released state) with the same name 548
as the file fragment received. Step 7102 then proceeds to step 7104
to determine whether such an incomplete file exists in the RAM
directory table 523. If found, step 7104 proceeds to step 7106,
which compares the header parameters of the fragment with those of
the existing file. If an unreleased file is not found, step 7104
proceeds to step 7134 to search the RAM directory table 523 for a
complete file in the released state with the same name as the file
fragment received.
[0112] Step 7106 then proceeds to step 7108, to determine whether
the header parameters (i.e. 542 to 552) of the fragment received
and the existing file match. If so, step 7108 proceeds to step
7110, which checks whether the file is completed. Otherwise, step
7108 proceeds to step 7116, which determines whether the file is
obsolete by checking whether the release date 543 of the fragment
received is newer than that of the existing incomplete file.
[0113] If the file is determined as complete in step 7110, step
7110 proceeds to step 7148, to end process 7100. Otherwise, step
7110 proceeds to step 7112 which determines whether the existing
unreleased file is in the write-protected section of flash memory
120. If step 7112 determined that the file is write-protected, step
7112 proceeds to step 7114, which determines whether the device is
in the initial configuration state which allows writes to the
write-protected section of flash memory. Otherwise, step 7112
proceeds to step 7150; a process which updates the incomplete file
in flash memory with the data received.
[0114] If step 7114 determines that the write-protected section is
writable, step 7114 proceeds to step 7150, to update the incomplete
file in write-protected flash memory. Otherwise, step 7114 proceeds
to step 7148, to end process 7100, with the file fragment received
being discarded, as the write-protected section 410 of the flash
memory 120 is read-only.
[0115] If step 7116 determines that the fragment is newer, step
7116 proceeds to step 7118, which checks whether either the file
fragment has a read-only access flag (e.g. 547), indicating that it
should be stored in the write-protected area, or that the existing
file is in the read-only section of flash memory 120. Step 7118
ensures a file of the same name may not exist in both
write-protected and read-write areas of the flash memory.
[0116] If step 7118 indicates neither the fragment nor the existing
file is read-only access, step 7118 proceeds to step 7122, which
deletes the existing incomplete file as it has been made obsolete
by the newer file fragment. Otherwise, step 7118 proceeds to step
7120 which checks whether write access to the write-protected area
of the flash memory system is available. If so, step 7120 proceeds
to step 7122 to delete the existing incomplete file by setting the
delete flag in the corresponding entry in the directory table 523.
Otherwise, step 7120 proceeds to step 7148, to end process 7100.
Step 7122 proceeds to step 7124, which attempts to search for a
free directory header in the appropriate write-protected 410 or
read-write 420 areas of the flash memory depending on the read-only
access flag in the received file fragment. The result of the search
in step 7124 is checked in step 7126. If the search located a free
header step 7126 proceeds to 7128 to initialize the fields (e.g.
542 to 552) in a new directory header 512 with the header
information from the received file fragment, and the local time
field 541 with the current time maintained by the clock task 225.
Otherwise, step 7128 proceeds to step 7130 to record a failure to
locate a free header, which proceeds to step 7132 to indicate a
file system failure.
[0117] Step 7128 also flags the status of the new file as
incomplete in the RAM directory table 523, and proceeds to step
7150 to commence storing the fragment as a new file.
[0118] Based on the search performed at step 7134, step 7136
determines whether a released file with the same name as the file
fragment received exists in memory 120. If no such file exists,
step 7136 proceeds to step 7144. However, if a released file is
found, step 7136 proceeds to step 7138 to compare the release dates
of the file and the file fragment. If the file fragment is newer,
step 7138, proceeds to step 7140 to check whether the device is in
an initial configuration state. If so, step 7140 proceeds to step
7124 commence storing the file fragment as a new file. Otherwise
step 7140 proceeds to step 7142, which checks whether either the
file fragment has a read-only access flag (e.g. 547), indicating
that it should be stored in the write-protected area, or that the
existing file is in the write-protected section of flash memory
120. If step 7142 indicates neither the fragment nor the existing
file is read-only access, step 7142 proceeds to step 7124 to
commence storing the file fragment as a new file. Otherwise, step
7142 proceeds to step 7148 to end process 7100.
[0119] Step 7144 determines whether the data in the file fragment
received belongs to a file intended for storage in the
write-protected area of memory 120 (e.g. based on access mode data
stored in the file fragment). If so, step 7144 proceeds to step
7146 to determine whether the unit is in an initial configuration
state, which allows write-access to the write-protected section 410
of the flash memory. Otherwise, step 7144 proceeds to step 7124 to
commence storing the file fragment as a new file. If step 7146
determines the unit is in initial configuration, step 7146 proceeds
to step 7124 to commence storing the fragment as a new file.
Otherwise, step 7142 proceeds to step 7148 to end process 7100.
[0120] The process 7150 to update an incomplete file with the data
quanta received in a file fragment is shown in FIG. 16. The update
quanta process begins at step 7152 by retrieving the pointer to the
page 530 in the data allocation unit where the data contained in
the file fragment is destined to be written. The data allocation
unit is identified by indexing the table of data allocation unit
pointers (e.g. 514) in the file's directory entry. The index is
calculated by using the starting quanta number (e.g. 741) in the
file fragment received, the quanta size (e.g. 550) and the number
of quanta held by each data allocation unit.
[0121] It is assumed quanta in the file fragment do not overlap
pages 530 in data allocation units, however it is straightforward
to enhance the procedure to cater for file fragments that contain
data that may overlap pages.
[0122] Step 7152 proceeds to step 7154, which determines whether a
data allocation unit identified by step 7152 has been allocated in
the data allocation unit table (e,g, 514) in the file's directory
entry. If so, step 7154 proceeds to step 7156, which determines if
the page identified in step 7152 contains all data quanta present
in the file fragment received. Otherwise, step 7154 proceeds to
step 7164, which searches the relevant pool of allocation units
(e.g. 521 or 522) for a free data allocation unit depending on the
read-only access flag of the file (e.g. 547). If step 7156
determines all data in the fragment are present in the file, step
7158, following step 7156, proceeds to step 7162 to end the update
quanta process 7150. Otherwise, step 7158 proceeds to step 7160 to
write the data in the fragment to the page and end the update
quanta process 7150 at step 7162. Step 7164 proceeds to step 7166,
which determines whether the search for a free allocation unit was
successful. If so, step 7166 proceeds to step 7168, which adds a
pointer (e.g 509) to the free allocation unit to the index in the
file's header table (e.g. 514), and proceeds to step 7160 to write
the fragment to the allocation unit. Otherwise, step 7166 proceeds
to step 7170, which records an error, and proceeds to 7172 to end
the update quanta process 7150 with a failure.
[0123] FIG. 17 shows the client device 100 receiving files from the
server transmission system as file fragments 2501, 2502 and 2503.
The file fragments 2501, 2502 and 2503 may be sent to the client
device 100 via a different communications interface 2506, 2508,
2510 (e.g. interfaces for a paging network, a wireless network, the
Internet, or a direct connection with the client device 100). The
file fragments 2501, 2502 and 2503 do not have to be sent or
received in any particular sequence. When the client device 100
receives all the file fragments 2501, 2502 and 2503, the device 100
reorganises all the fragment into their proper sequence for
processing (e.g. based on the fragment identifier for each file
fragment 2501, 2502 and 2503 indicating their sequence).
[0124] Similarly, commands and configuration data can be sent to
the client device 100 via a different communications
interfaces.
[0125] The file system also implements erasure correction recovery
of missing fragments. All files have an associated 32-bit cyclic
redundancy check (CRC) and a CRC calculated locally and stored for
every page to check the integrity of the data stored in the flash
memory page. These erasure correction pages follow the end of the
file. A file missing core fragments also has provision for Erasure
Error Correction (EEC) data at the end, and when the necessary file
data is received or able to be reconstructed using the EEC data,
any EEC data is freed and file marked as "complete".
[0126] Once the file is complete, it cannot be altered or patched
by the VM task 213, but can only be deleted by the flash file
system.
[0127] The flash file system supports a locking mechanism to
prevent them from being deleted during a garbage collection process
performed by the file system task, particularly if a new version of
the file is available. A locked file gives an application
guaranteed access to that file. The flash file system always locks
existing files on initialisation and newly completed files. Files
can only be unlocked by the application.
[0128] The file system housekeeping process 7200 is shown in FIG.
18. Once the flash file system is started at step 7202, process
7200 executes a number of procedures in turn to remove incomplete
files (at step 7204), remove deleted files (at step 7206) and to
check completed files for deletion (at step 7208). The function of
steps 7204, 7206 and 7208 are described in further detail with
reference to processes 7300, 7400 and 7500 respectively, as
depicted in FIGS. 19, 20 and 21. The actions of determining which
file to delete and actually deleting them are made separate by the
use of a deletion flag to indicate to the deletion process that the
file is a delete candidate. This allows the relatively slower
deletion process shown in FIG. 20 to be handled by the housekeeping
task 223 thereby speeding up any process which wishes to delete a
file.
[0129] FIG. 19 is a flow diagram of the process 7300 for removing
incomplete files if they are not completed within a specified time.
If the number and type of free allocation units (i.e
write-protected or read-write), related to the access mode of the
file, fall below a preset threshold less time is allowed for
incomplete files to become complete to free up more allocation
units. Process 7300 begins at step 7304 by determining whether the
Real Time Clock (RTC) of the microprocessor 102 is returning a
valid time value. If the RTC output is not a valid time value,
process 7300 ends and execution returns to process 7200. Otherwise,
if the RTC output is valid, step 7304 proceeds to step 7306 to
locate a first directory entry in the flash memory 120 for
processing. Step 7306 proceeds to step 7308 to retrieve data stored
in the directory entry and determine whether the directory entry
corresponds to a file. If the directory entry does not correspond
to a file, step 7308 proceeds to step 7330. Otherwise, step 7308
proceeds to step 7310 to determine whether the file has been marked
for deletion. If the file has been marked for deletion step 7310
proceeds to step 7330. Otherwise, step 7310 proceeds to step 7312
to determine if the file is complete. If so, step 7312 proceeds to
step 7330. Otherwise, step 7312 proceeds to step 7314 to determine
if the file is stored in the write-protected section of the flash
memory 120 (i.e. read-only memory if not in initial
configuration).
[0130] If the file is stored in write-protected memory, step 7314
proceeds to step 7324 to determine if there are sufficient free
allocation units in the read-only memory for storing more data. If
the write-protected memory has sufficient capacity, step 7324
proceeds to step 7330. Otherwise, step 7326 generates the time
difference between the local creation date of the file (i.e. the
time when the first portion of the file was stored in memory 120)
and the time at which the most recent portion of the file was
received. Then, step 7328 determines whether the time difference
generated at step 7326 exceeds the predetermined timeout value for
the write-protected memory. If the generated time difference
exceeds the timeout value for the write-protected memory, then step
7328 proceeds to step 7330. Otherwise, step 7328 proceeds to step
7322 to mark the file for deletion.
[0131] If the file is stored in the read-write section of the flash
memory 120 (i.e. read-write memory), then step 7314 proceeds to
step 7316 to determine if there are sufficient free allocation
units in the read-write memory for storing more data. If the
read-write memory has sufficient capacity, step 7316 proceeds to
step 7330. Otherwise, step 7318 generates the time difference
between the local creation date of the file (i.e. the time when the
first portion of the file was stored in memory 120) and the time at
which the most recent portion of the file was received. Then, step
7320 determines whether the time difference generated at step 7318
exceeds the predetermined timeout value for the read-write memory.
If the generated time difference exceeds the timeout value for the
read-write memory, then step 7320 proceeds to step 7330. Otherwise,
step 7328 proceeds to step 7322 to mark the file for deletion.
[0132] Step 7322 marks the file for deletion (e.g. by setting the
deletion flag for the file) and proceeds to step 7330 to determine
whether all directory entries in the memory 120 have been
processed. If so, process 7300 ends and execution returns to
process 1200. Otherwise, step 7330 proceeds to step 7308, where a
different directory entry is selected and processed.
[0133] FIG. 20 is a flow diagram of the process 7400 for removing
files marked for deletion from the memory 120. Process 7400 begins
at step 7404 by locating a first directory entry in the flash
memory 120 for processing. Step 7404 proceeds to step 7406 to
retrieve data stored in the directory entry and determine whether
the directory entry corresponds to a file. If the directory entry
does not correspond to a file, step 7406 proceeds to step 7416.
Otherwise, step 7406 proceeds to step 7408 to determine whether the
file has been marked for deletion by examining the file's status in
the directory table 523. If the file has not been marked for
deletion step 7408 proceeds to step 7416. Otherwise, step 7408
proceeds to step 7410 to determine if the file is open (e.g.
currently being opened or used). If so, step 7410 proceeds to step
7416. Otherwise, step 7410 proceeds to step 7412 to determine if
the file is locked (e.g. locked so it will not be deleted). If so,
step 7412 proceeds to step 7416. Otherwise, step 7412 proceeds to
step 7414 to delete the file. Step 7414 proceeds to step 7416 to
determine whether all directory entries in the memory 120 have been
processed. If so, process 7400 ends and execution returns to
process 7200. Otherwise, step 7416 proceeds to step 7406, where a
different directory entry is selected and processed. The main file
system housekeeping loop is also responsible for checking
incomplete files and building any missing fragments using received
erasure correction fragments where possible, as depicted in FIG.
21. A CRC is performed on any newly completed files to check their
integrity. The process of assembling files is decoupled from the
receive fragment process shown in FIG. 15, to simplify both.
[0134] FIG. 21 is a flow diagram of the process 7500 for marking
incomplete files for deletion from the memory 120. Process 7500
begins at step 7504 by locating a first directory entry in the
flash memory 120 for processing. Step 7504 proceeds to step 7506 to
retrieve data stored in the directory entry and determine whether
the directory entry corresponds to a file. If the directory entry
does not correspond to a file, step 7506 proceeds to step 7530.
Step 7530 determines whether all the directory entries in the
memory 120 have processed. If so, process 7500 ends and execution
returns to process 7200. Otherwise, step 7530 selects another
directory entry from memory 120 and proceeds to step 7506 to
process the new directory entry.
[0135] If step 7506 determines that the directory entry corresponds
to a file, step 7506 proceeds to step 7508 to determine whether the
file has been marked for deletion. If so, step 7508 proceeds to
step 7530. Otherwise, step 7508 proceeds to step 7510 to determine
whether the file is complete. If the file is complete, step 7510
proceeds to step 7530. Otherwise, step 7510 proceeds to step 7512
to generate the core quanta present in the file (i.e. the amount of
data for the file, excluding the erasure correction fragments) and
the core quanta still required (i.e. the amount of data still
required for the file to be complete). The core quanta required may
be determined, for example, by subtracting the core quanta present
in the file from the file length. Step 7512 proceeds to step 7514
to verify whether the quanta present for the file (determined at
step 7512) is actually stored in the file system. If so, step 7514
proceeds to step 7518. Otherwise, step 7514 proceeds to step 7516
to attempt to record core quanta of the file using the error
correction fragments. Step 7516 then proceeds to step 7518.
[0136] Step 7518 determines whether the core quanta for the file
has been stored in memory 120 (e.g. including the fragments
recovered at step 7516). If not, step 7518 proceeds to step 7530.
However, if all the core quanta is present, step 7518 proceeds to
step 7520 to determine whether the file has now been marked for
deletion. If so, step 7520 proceeds to step 7530. Otherwise, step
7520 proceeds to step 7522 to generate CRC data for the file. Then,
at step 7524, the CRC data generated at step 7522 is compared with
the CRC data stored in the header of the file. If the CRC data
match, step 7524 proceeds to step 7526 to mark the file as
completed, and truncates the erasure correction fragments for the
file. Otherwise, step 7524 proceeds to step 7528 to mark the file
for deletion. Both steps 7526 and 7528 proceed to step 7530.
Virtual Machine
[0137] The virtual machine (VM) 321 is a software module executed
by the processor core 102 as a task, VM task 221, which runs
applications stored in the flash file system. The VM is implemented
as a byte-code interpreter that allows a single application to be
run at any time, as the interpreter is single threaded. The VM may
include, for example, two separate program counters and two
separate stacks in order to run multiple applications. The use of
byte-code allows the execution of applications from the data-flash
memory that otherwise would not be able to be run by the processor
102.
[0138] A VM application is primarily event driven. It polls and
acts on an event queue which accepts asynchronous events from user
I/O (e.g. buttons), scheduled timer events, etc.
[0139] A particular application normally resides in a single
identified file, as discussed below, but the interpreter supports
call and return of functions in other files, which allows the
application code to be maintained in a modular fashion, and thereby
reducing file size downloads.
[0140] At any time there may be many application and data files
resident in the flash file system. The following process determines
what application to run and also manages the removal of redundant
files.
[0141] The VM 321 when started in the VM task 221 first attempts to
run a file with name "0" (i.e. the "init" application 611, as shown
in FIG. 9) typically stored in the write-protected area 410 of the
flash file system 400. The init application 611 reads a
configuration "init-tab" table file 621 which determines what file
or files correspond to applications to be executed, the data files
required for each application and the conditions under which the
applications should each be executed (e.g. by using the pointers
631 to those files stored in the init-tab file 621). The init-tab
file 621 is stored in the read-write area 420 so it can be
regularly updated as the application and data files 622, 623, 624
stored in the flash memory 120 are updated. The init application
611 is also responsible for identifying and flagging files 625 that
are candidates for deletion to help ensure the flash file system
always has space for new files. Any files not identified in
init-tab file 621 are delete candidates.
TABLE-US-00002 TABLE 1 Event Application Data Files A 12 2, 3, 7,
8, 9, 13 . . . A 6 2, 3, 5, 7, 8, 9, 10, 11 A, B 1 2, 3, 4, 5
[0142] Table 1 provides an example of the data in an init-tab file
621 which refers to three applications, respectively stored in
files 12, 6 and 1. The first column identifies the events from the
event queue that may be used to trigger the running of the
application file in the second column. The last column determines
data files required by each application.
[0143] The init-tab file 621, as shown in Table 1, states that an
initial version of the application governing the operation of the
device 100 was placed in file "1", and used data files 2, 3, 4 and
5. After some time, a new version of the application was produced
with expanded functionality, and transmitted. This new version of
the application used all files that original version used, with the
addition of new data files 7, 8, 9, 10 and 11. The new version of
the application was placed in file 6. Finally, the latest version
of the application was created and transmitted to the device 100.
The latest version of the application eliminated several of the old
data files (4, 5, 10 and 11) and added new data file 13. The latest
application file has been placed in file 12. The init-tab file 621
endeavours to run application file 12 if possible, otherwise
application file 6 is run, or otherwise application file 1. An
application file is only run if the init-tab file 621 is able to
verify that all required data files for execution of a particular
application file are available.
[0144] The init-tab file may also be structured to identify
additional files for each event that need not be present to run the
corresponding application, but must not be identified as delete
candidates. This allows optional files to be present to be used by
the application, if desired, without them being removed by the init
application 611. If the space for new files or the number of free
directory entries is below a certain threshold, file system
resource may be recovered by removing one or more sets of files in
init-tab corresponding to the lower priority events depending on
the amount of resource to be recovered.
[0145] New data files 622, 623, 624 for an application may be
received using receive process 1100 in FIG. 15, and once received,
the files are locked by the application (e.g. by setting a
read-lock flag corresponding to each file). Such locking prevents
that application from using other data files which are not part of
the current version being used by the application (e.g. files for a
previous version). The application may use data files for a
previous version by first unlocking the files for the current
version being used, and then locking the files for the version
selected for use.
[0146] The byte-code instructions of an application can be read
into a small cache maintained in the RAM 108 to reduce the flash
access overhead. The cache size being empirically set to balance
cache load overhead with cache misses. The VM implements a data
stack to store the local temporary variables and parameters. The
design of the byte-code instructions used by the application files
and interpreted by the VM are orthogonal to ease compiler design
whilst minimizing the number of instructions. They are grouped into
8-bit, 16-bit, 32-bit, floating-point groups and by number of
operands for execution efficiency. A common set of operations is
supported in all arithmetic groups (such as arithmetic, memory
operations etc.), and additional operations may be supported for
each group (e.g. bit tests for integers or for floating-point
numbers etc.). Other operations, which are not part of the VM (such
as trigonometric operations) are implemented, as desired, by
function calls to specific files. Hardware or software resources
are accessed using system calls, such as flash read/write. The
function and system calls are executed using VM instructions
involving unique function and system operational codes,
respectively. A single operand is accepted to determine the
operation to be performed, and any required parameters are passed
via the data stack. The system calls are handled externally of the
VM and this allows extensions that do not affect the core of the
virtual machine. Calls to the flash file system, for example, allow
directory information and file data information to be read.
[0147] An application can make a system call to the VM to reboot
the device 100, if this is necessary for the application to now be
run or take advantage of any new OS components sent to the device
100.
Air Protocol
[0148] The receiver 110 and the processor 102 utilize an air
protocol to receive information including commands, time of day and
file fragments. Information is transmitted using the air protocol
by a server transmission system that has a server communications
application to implement the air protocol and which can use an
existing pager communications network infrastructure for
transmission. The air task 214 controls communications between the
processor 102 and the receiver 110. The air protocol is a
communications protocol that, in accordance with the OSI layer
model, has different attributes for a physical layer, a data link
layer, a network layer and a transport layer
[0149] The device 100 can use an existing communications network
infrastructure to receive information, for example a paging
network. Accordingly the link layer of the air protocol can be
tailored to support transmission of text messages as defined by the
Post Office Code Standardisation Advisory Group (POCSAG)
standard.
[0150] The physical layer is simply the bit stream that is
broadcast and received by the receiver 110. All the receivers 110
of different client devices 100 are normally tuned to the same
carrier frequency for an application. This allows all of the client
devices 100 to be updated with application files, as required. A
unique ID or address for each device 100 (such as the serial number
for the device 100), is available to filter the packets for an
individual device 100. The device 100 is therefore identified by a
paging address, and transmissions can be made for a number of
paging addresses. The link layer supports a sub-set of the
printable 7-bit characters, as discussed below.
[0151] The network layer supports the transmission of multicast
packet data. The packets include address information to send the
data to a single device address, a group of devices or all devices.
A type parameter is used to identify the data carried, and the data
types include commands, time and file segments, ie fragments. The
network layer also enables data authentication and validation.
[0152] The transport layer supports the carriage of an arbitrary
number of data octets of any length up to a maximum length
supported by the link layer, which for the device 100 is 170
octets. The transport layer also provides error detection and
optional correction so that corrupted packets can be identified and
corrected
[0153] For the transport layer, the data is pre-pended with its
length in octets. The length is represented by an octet
representing an unsigned 8-bit number. A CRC-32 is calculated from
this and appended in little endian format as a quad-octet. The
resulting message is then broken into quad-octets, padding with
zeroes as necessary. Each quad-octet is converted to group of 5
ASCII characters using a base 85 encoding scheme, similar to
RFC1924, as discussed below.
[0154] The quad-octets are each treated as a little endian 32-bit
unsigned number. To convert repeatedly they are divided by 85,
taking the remainders at each step. The 5 remainders are placed in
little endian order. They are each mapped to a printable character
by using a mapping scheme that has monotonically increasing values,
which allows for simpler decoding (ie without using a look-up
table).
[0155] The resulting ASCII message is then padded out to a multiple
of 20 characters by padding with `}`. A `}` is used because a group
of 5 is an illegal BASE-85 encoding. This is to ensure that the
message occupies a complete number of POCSAG codewords (each
contains 20-bits, 20 characters=7 codewords). The group is encoded
also using the base 85 encoding scheme.
[0156] For the network layer, an air protocol packet has the
following structure:
TABLE-US-00003 <packet> ::= <Protocol
Id><payload> <Protocol Id> ::= <octet>
<payload> depends on <protocol Id> which designates the
protocol type.
[0157] Protocol Id has a value that identifies the protocol as
follows.
[0158] <Protocol Id>=1 allows sending of a data of a
specified type. This can be sent to specified processors by the
parameter <Processors> if the device 100 has more than one
processor 102. More than one processor may be included in the
device 100 in situations where a GPS receiver of the device 100 has
a processor that is considered to be a separate processor for which
files may need to be downloaded. Alternatively, an additional
processor 102 may be included to be dedicated to a particular OS
task to enhance the cycle time for that task and in turn enhance
the speed of the device 100. Then [0159]
<payload>::=<Address><Processors><Type><Data
[0160] <Protocol Id>=2 allows transmission of a specified
data type to specified processors in a group of devices with a
timestamp. [0161]
<payload>::=<Address><Processors><Type><Time
stamp><Data>
[0162] <Protocol Id>=3 allows transmission of authenticated
data to specified processors with specified type to a group of
devices with a timestamp. [0163]
<payload>::=<Address><Processors><Type><Time
stamp><Authentication><Data>
[0164] <Protocol Id>=4 allows transmission of a specified
data type to specified processors in a group of devices with a
timestamp and a local time offset. [0165]
<payload>::=<Address><Processors><type><Time
stamp><Local time offset><Data>
[0166] <Protocol Id>=5 allows transmission of specified data
type to a set of units in a specified product and platform range
with timestamp and local time offset. [0167]
<payload>::=<Product
Id><Platform_range><Address><Type><Data>
[0168] <Protocol Id>=6 allows transmission of specified data
type to a set of units in a specified product and platform range
with timestamp. [0169] <payload>::=<Product
Id><Platform_range><Address><Type><Time
stamp><Data>
[0170] <Protocol Id>=7 allows transmission of specified data
type to a set of units in a specified product and platform range
with timestamp and authentication.
TABLE-US-00004 <payload>::=<Product
Id><Platform_range><Address><Type><Time
stamp><Authentication><Data>
[0171] <Protocol Id>=8 allows transmission of specified data
type to a set of units in a specified product and platform range
with timestamp and local time offset.
TABLE-US-00005 <payload>::=<Product
Id><Platform_range><Address><type><Time
stamp><Local time offset><Data>
[0172] For product Id:
[0173] The product Id allows the particular pager address (cap
code) to support multiple products by identifying which product
should receive the packet: [0174] <Product
Id>=<octet>
[0175] For platform range:
[0176] The platform range determines what releases of the platform
should receive the packet by means of a range of acceptable
platform build numbers:
TABLE-US-00006 <Platform_range> :: =
<platform_min><platform_max>
<platform_min>,<platform_max> = <platform> =
<octet_LSB> <octet_MSB>
[0177] For addressing:
TABLE-US-00007 <address> :: =
<broadcast>|<group>|<serial> =
<octet><octet><octet> <octet>
<broadcast> ::= 0xFF, 0xFF, 0xFF, 0xFF <group> :: =
0xFF,0xFF, <group Id> <group Id> ::=
<octet_LSB><octet_MSB> = 0x0000 .. 0xFFFE
<serial> :: =
<octet_LSB><octet><octet><octet_MSB>
[0178] Specifies all devices, a particular device or a group of
devices. [0179]
<Processors>::=<octet>=<processor1:lsb> . . .
<-processor8:msb>
[0180] Specifies a set of processors from a maximum of 8.
[0181] Message types specifies how the system should handle the
data being transmitted:
TABLE-US-00008 <type> ::= <octet> = <file
fragment> | <set time> | <set groups> | alarm>
< file fragment> ::= 1 <set time> ::= 2 <set
groups> ::= 3 <user data> :: = 4 <user alarm> :: =
5
[0182] The actual message data transmitted:
TABLE-US-00009 <data>::={<octet>}<file
fragment><set time><set groups><user
data><user alarm>
[0183] For a file fragment sent to a specified memory:
TABLE-US-00010 <file fragment> ::=
<header>{<byte>} <header> ::=
<CRC32><date><length><data><erasure><num-
ber> <access><name>
<size><count><correction> <CRC32> ::=
<LSB:octet><octet><octet><MSB:octet> =
CRC32 of core file contents <date> ::= <time> = Release
date <length> ::=
<LSB:octet><octet><octet>< MSB:octet> =
file length in bytes <data> ::= <LSB:octet
><MSB:octet > = length of all data quanta and erasure
correction quanta in bytes <erasure> ::= <LSB:octet
><MSB:octet > = length of all erasure correction quanta
only <number> ::= <LSB:octet ><MSB:octet > =
first quanta number in this fragment <access> ::=
<section:LSN><storage:MSN> <name> ::=
<byte> = filename <size> ::= <byte> = quanta size
<count> ::= <byte> = quanta count <correction>
::= <nibble:E><nibble:c> = erasure correction, E = k
and c = n-k <section> ::= <FLASH_RO> | <FLASH_RW>
= flash section type <FLASH_RW> ::= 0 = Read/write section
<FLASH_RO> ::= 1 = Read only section <storage> ::= 0 ..
15 = memory id to store fragment
[0184] The parameter set groups allow up to eight group addresses
to be set for a device 100 identified by its 10, ie its serial
number.
TABLE-US-00011 <set groups> ::= <group1:group> ..
<group8:group> <group> ::= <LSB:octet
><MSB:octet
[0185] The parameter set time sets time using a previous time
reference to reduce error.
TABLE-US-00012 <set time> ::= <time reference><time
correction><Local time offset> <time reference> :=
<time>, previous timestamp in payload <time correction>
::= <octet> = 0 .. 240 (seconds = 4 minutes)
[0186] The format of user data
TABLE-US-00013 <user data> ::=
{<offset><length><data>} <offset> ::= 16
bit unsigned <length> ::= 16 bit unsigned <data> ::=
<length> octets.
[0187] For user alarms to be set:
TABLE-US-00014 <user alarm> ::= {<set ualarm> |
<cancel ualarm>} <set ualarm> ::= `S`<ualarm
number><ualarm record> <cancel ualarm> ::= `C`
<ualarm number> <ualarm number> ::= <octet> = 0
.. 3 (device may use less than this)
[0188] A <ualarm record> is large enough for 4 <time>s
however only 2 are defined:
[0189] And the rest may or may not be used for <time>s in the
future, they may be just data.
TABLE-US-00015 <ualarm record> ::= <ualarm
time><ualarm useby><ualarm 0><ualarm 0>
<ualarm time> ::= <time> (the alarm is to go off)
<ualarm useby> ::= <time> (the time the alarm is not to
be attempted after) <ualarm 0> ::= a zero of number of octets
in <time>.
[0190] The time format, which defines days, minutes, seconds since
1 Jan. 2000, GMT is:
TABLE-US-00016 <time> ::=
<octet_LSB><octet><octet><octet_MSB> =
<dddddddddddddddmmmmmmmmmmmssssss>
[0191] days most significant 15 bits, minutes next 11 bits, seconds
least significant 6 bits [0192] <Time
stamp>::=<time><Local time
offset>::=<octet>
[0193] specifies signed difference in half-hour increments between
local and UTC time.
[0194] E.g. 20=+10 hours
[0195] To supply an authentication keyword: [0204] [0196]
<authentication>::=<octet_LSB><octet><octet><o-
-ctet_MSB>
Command Processor
[0197] The command processor task 222 and associated module 322
implements a command line interface for a computer attached via the
asynchronous serial interface 115.
[0198] For example a "LS" command would produce, as shown in the
example Listing 1 below, first a header output and then a listing
of the files of the device 100 written out one file per line. The
fields in the lines correspond to the fields in the header, and
what each field means is described in Table 2 below:
TABLE-US-00017 TABLE 2 Field Description Na Directory name, the
name in the directory written out as a <byte2> since the file
may be incomplete and so that is not printable. N Name of the file
written as a <graphicchar>. If the name is not a graphic char
then the "~" is written instead so that the columns stay correctly
spaced. Not "~" is 0xxx so it may actually be the name of the file.
DN Directory number of the file written out in <byte8>. TY
File type written out in <byte8>. Length This is the length
of the file in bytes and the value is written out in <byte8>.
CRC32 This is the CRC32 (32 bit CRC) of the file and its value is
written out as <byte8>. R Date This is the Release Date of
the file written out in <time>. C Date This is the Create
Date of the file written out in <time>. DQ The number of Data
Quanta for the file in <byte4>. EQ The number of Erasure
Quanta for the file in <byte4>. QS The Quanta Size for the
file in <byte2>. QC The Quanta Count for the file in
<byte2>. EC The Erasure Correction byte for the file in
<byte2>. A The Access for the file .written as a
<char>. A "W" means it is read and writeable on the media and
"R" means that it is read only on the media. FG The flag byte which
flags various information about the file. The value is written out
as <byte2> and the value is the sum of various flags. Their
values are: Deleted: 0x01 Incomplete: 0x02 Complete: 0x04 Released:
0x08 Locked: 0x10 Open: 0x20
TABLE-US-00018 Listing 1 Na N DN TY Length CRC32 R_Date C_Date DQ
EQ QS QC EC A FG >ls 51 Q 00 00 000004E6 12B0DD16 0F922003
0F926EC0 000A 0000 80 01 10 W 38 4D M 01 00 00002A4C D2E930D3
0F921FDC 0F926E95 0055 0000 80 01 10 W 38 00 ~ 02 00 000004F6
FD16F9E8 0F92689E 0F926E91 000A 0000 80 01 10 W 08 31 1 03 00
00006950 76BED359 0F92689E 0F926E28 00D3 0000 80 01 10 W 38 32 2 04
00 00006D08 CD49541C 0F92689E 0F926DC0 00DB 0000 80 01 10 W 08 33 3
05 00 0000736A 23BCF15C 0F92689E 0F926D4E 00E7 0000 80 01 10 W 08
3A : 06 00 00000B4B 44658709 0F92689E 0F926D43 0017 0000 80 01 10 W
38 40 @ 07 00 0000001E 7A09C6EA 0F92689E 0F926D43 0001 0000 80 01
10 W 18 41 A 08 00 00002A0C 58EC9718 0F92689E 0F926D18 0055 0000 80
01 10 W 08 5F .sub.-- 09 00 00000001 D202EF8D 0F92689E 0F926D17
0001 0000 80 01 10 W 18 4B K 0A 00 0000EA60 944B18FD 0F9261E0
0F9468AE 01D5 0000 80 01 10 W 02 >>ls Q 51 Q 00 00 000004E6
12B0DD16 0F922003 0F926EC0 000A 0000 80 01 10 W 38
2. Transmission System Infrastructure
2.1 File Transmission
[0199] The files to be transmitted to the device 100 are each first
broken up into a size that can be encapsulated by the air protocol
and sent using a server transmission system, described below and
shown in FIG. 23. The file fragments produced in accordance with
the air protocol can also be transmitted to the device 100 using a
remote relay unit (RRU) 1900, as described below and shown in FIG.
22.
[0200] The file to be transmitted 701, as shown in FIG. 10, is
treated as an arbitrary sequence of bytes, 710, up to the limit
supported by the flash file system described above. Each byte, 711
is an 8-bit binary value. Consecutive sequences of bytes 721 are
assembled into equal length groups of quanta 720. The last quanta
722 to be transmitted is padded with zeros, as required, to ensure
that the last quanta 722 has the same size as the other quanta
(e.g. 720) being transmitted. Quanta are all equal length for a
particular file and are numbered consecutively starting at
zero.
[0201] A file fragment 730 suitable for transmission consists of a
header 742 followed by data 743. The data is a sequence of quanta
720. The data in the last fragment is padded with zeros as
required.
[0202] The header 742 contains information related to the data 743
in a file fragment 730, and consists of a number of parameters
(described as follows) transmitted in little endian order for
compatibility with the representation used by the firmware compiler
2002. Each header 742 starts with a 32-bit cyclic redundancy check
731 calculated for the data 743 contained in the file fragment 730.
This is followed by a 32-bit date/time stamp 732 for the file 701
in the format described in the air protocol section above, and a
32-bit length 733 representing the total length of the file 701
(e.g. in bytes), both of which may be derived from the file's 701
directory information. This is followed with a 16-bit count of the
total number of data quanta 734 required to send the whole file 701
(i.e the core data quanta, or the data corresponding to the actual
data in the file 701), and the total number of erasure correction
quanta 735 generated for the file 701, as described below. A 16-bit
quanta number 736 identifies the number assigned to the data quanta
that first appears in the current file fragment 730 being
transmitted. If the quanta number 736 in the current file fragment
730 is greater than the number of data quanta required to send the
whole file 701 (i.e. as represented by 734), then the data 743 in
the fragment 730 is erasure correction data. An 8-bit access field
736 determines if the file 701 is to be placed into read/write 420
or write-protected 410 section of the file flash system 400, based
on the value represented by the access field 737 (e.g. as "0" for
read/write access and "1" for write-protected access). Files (e.g.
701) are given a single 8-bit name 737 (e.g. an 8-bit binary
number) of which typically the last 255 values represented by the
8-bits are available for use. The number of bytes per quanta size
and the number of quanta per fragment are represented by 8-bit
values 738 and 739, respectively. The correction information 741
represents an 8-bit value and identifies the level of erasure
correction used, the nibbles at the upper and/or lower nibble
locations, representing as k and n-k respectively (where k is an
integer with a predetermined value and n is the total number of
fragments in the group). For a particular file all parameters in
the header are held constant with exception of the quanta number
736 which corresponds to the file location of the data being
transmitted
[0203] File fragments may be transmitted in any order and there is
a benefit to be gained by interleaving fragments to improve system
performance if a number of consecutive fragments are lost by the
transmission system. The system also allows a file fragment to be
transmitted repetitively to increase the probability the complete
file is received.
[0204] If desired, erasure correction fragments (ECFs) can be
additionally generated from the fragments derived from the file
data. ECFs allows lost file fragments to be recovered more
efficiently than relying on multiple transmissions of the sample
fragment. This redundancy greatly improves reception performance
under adverse signal conditions. The level of redundancy is set on
a file by file basis depending on the reliability desired based on
transmission loss, number of fragments, etc
[0205] FIG. 11 is a schematic diagram showing how erasure
correction fragments are generated. A set of k consecutive
fragments (e.g. 810, 820), are used to generate one or more (ie
n-k) erasure correction fragments 840 using an ECF process
performed by an ECF module 830. In addition to the data in the sets
of fragments 810, 820 received by the module 830, the module 830
accepts further input for determining the number of k groups of
fragments 831 in each set and the desired number of erasure
correction fragments 832 to be generated based on the groups of
fragments in each set. The size of the erasure correction fragment
840 is the same size as the original fragments as received by the
ECF module 830.
[0206] The ECF module 830 generates the desired erasure correction
fragment using finite field arithmetic on a nibble-by-nibble basis.
For example, lower nibbles from the same locations from each of the
original fragments in each set (e.g. nibbles 811 to 821) are each
used to generate a single corresponding erasure correction nibble
(e.g. 841). Similarly, the upper nibbles at locations (e.g. 812 to
822) are used to generate a corresponding upper erasure correction
nibble (e.g, 842). A field size of 16 is therefore sufficient and
means an efficient implementation of the finite field arithmetic
using 8-bit arithmetic natively supported by the processor 150.
[0207] The ECF process may use a Lagrange polynomial function to
generate erasure correction data. For example, each integer value
of x (along the x-axis, as shown in FIG. 12) corresponds to a
nibble in their order of appearance in each set (e.g. 811 to 821).
The value on the y-axis of FIG. 12 at each x-location corresponds
to a numeric value represented by each nibble respectively. The
polynomial coefficients of the Lagrange function are determined
such that the function uses the data values of the nibbles at their
equivalent x-locations 901, 902, 903, 904, as depicted in FIG. 12
(i.e. the coefficients define an approximated function that passes
through the y-axis value corresponding to each x-location 901, 902,
903, 904, etc.). The erasure correction data may also be generated
by evaluating the polynomial at other points 905, 906, 907 as
required. The finite field arithmetic limits the sum of the number
of erasure correction fragments and original fragments to less or
equal to 16.
[0208] In the client processor device 100, the Lagrange polynomial
interpolation may be used in a similar fashion to retrieve any
missing fragments provided that at least a unique k (as defined by
the correction information 741) of the original number of n
transmitted file fragments 730 have been received.
2.2 Web Server System
[0209] A user is able to control one or more client devices 100,
and these can be allocated to an identified group using a group Id,
as discussed above for the air protocol.
[0210] A web server system 1005, as discussed below, is provided
for users to access and enable them to: [0211] (i) Create a group;
[0212] (ii) Change a group's details; [0213] (iii) Change devices
in a group; [0214] (iv) Remove a group; [0215] (iv) Configure a
group of devices 100; [0216] (v) Display data received from one or
more devices 100.
2.3 Transmission System
[0217] The infrastructure includes a server transmission system, as
shown in FIG. 23, having a file collator 1005, an update manager
1006, a master database 1007, performance monitor 1008 and a
wireless transmission apparatus 1010. The server transmission
system also provides a customer web interface 1009 via a web server
provided by the file collator 1005.
[0218] The file collator 1005 (or central transmission server)
loads and maintains information that controls the operation of the
device 100, and to create and store instruction and data files used
by that device. The data is stored in a relational database 1007,
and data file suitable for transmission to the device are created
by a file formatter 1020 and kept on disk in a currently
transmittable file storage database 1021 from where they can be
retrieved by an update scheduler 1025 as required. Instruction
files used by the device 100 are also stored in the currently
transmittable file storage database 1021. The file collator 1005
includes a router application manager 1017, a configuration manager
1018, a file loader 1019 and a file formatter 1020, in addition to
the currently transmittable files storage database 1021.
[0219] The master database 1007 interfaces with a billing module
1011, which enables the data files for transmission to the devices
to be updated with a range of billing data. For example, the master
database 1007 may retrieve call tariff rates for different
carriers, carrier rate structure (e.g. peak and off peak rates),
carrier billing data and customer billing data via the billing
module 1011 for storage in the master database 1007
[0220] The router application manager 1017 generates files for
transmission to the device 100. Modifications to the manner in
which a task is performed by the device 100 are achieved by sending
a new version of the relevant file to the device 100. The file is
written directly to the current transmittable file storage database
1021 from the file formatter 1020. From this database 1021, the
update scheduler 1025 forming part of the update manager 1006,
retrieves the file for transmission at an appropriate time.
[0221] The configuration manager 1018 enables manual changes to be
made to the information maintained in the master database 1007. The
configuration manager 1018 includes a number of screen layouts that
can be used to review and change a different section of the
database 1007. Each user of the system is issued a user name and
password that determines which screen the user can access. Screens
that are used to maintain a particular client's information are
accessible only to that client, using the Internet 1022, and a
suitable web browser constituting the customer web interface 1009,
and authorized company personnel. The configuration manager 1018
also interfaces with the internal database maintenance module 1032,
which maintains the consistency of the data in the master database
1007 and may provide the configuration manager 1018 with an updated
view of the database 1007.
[0222] The configuration manager 1018 includes two sub modules (not
shown in FIG. 23), namely a transmission manager and a client web
interface. The transmission manager enables user control of the
schedule for transmission of files maintained in the currently
transmittable file storage database 1021 to the device 100. Files
can be transmitted regularly either at a set interval or at a
specific day and/or time, or a file can be transmitted on a one off
basis. Files can be assigned a priority when transmitted. The
client web interface enables users to maintain personal information
and device configuration information, as discussed previously.
[0223] The file loader 1019 provides a consistent process by which
data contained in externally sourced files is loaded into the
master database 1007 (via file formatter 1020), and is tested and
released to production in a controlled manner. Given the differing
nature of the various files that need to be loaded, the file loader
1019 is modular in nature so that each type of file can be handled
individually when necessary, whilst assuring the same overall
process is followed. Externally sourced files, such as structure
data files 1023, and external entity information files 1024, are
shown in FIG. 23, and may be provided by an infrastructure provider
or service provider. The information files may include costing
information and other data.
[0224] The file loader 1019 ensures that the following steps are
undertaken during the loading of a file:
[0225] (1) Parsing. Errors in file formats are detected and
handled.
[0226] (2) Preliminary load. The data is first loaded into a
"staging" area within the database so that it can be verified.
[0227] (3) Redundancy checks. New data is compared against data
maintained in the master database 1007, if there is no difference,
the new data is discarded. If differences are found however, the
nature of the differences are gauged so that appropriate levels of
testing can be performed.
[0228] (4) Testing. The new data is tested in an appropriate
fashion.
[0229] (5) Approval. Authorised personnel check the results of the
previous stages and approve, or not, the new data as suitable for
production.
[0230] (6) Release. The new data is released to production and the
file formatter 1020 is instructed to regenerate the affected files
in the current transmittable file storage database 1021. This
typically involves scheduling one or more files for transmission to
devices 100, depending upon the importance of the new data.
[0231] The file formatter 1020 provides a consistent framework for
the generation of files that are required for operation of the
device 100. Similar to the file loader 1019, the file formatter
1020 provides a consistent structure for the files whilst still
allowing for the individual requirements of each type of file. The
file formatter 1020 places the generated files in the current
transmittable file storage database 1021 where the files are
accessible by the process responsible for initiating the
transmission of the files (e.g. the update scheduler 1025). The
current transmittable file storage database 1021 forms part of the
directory hierarchy on a main server's disk which is accessible to
all processes involved in the generation and transmission of files,
namely the router application manager 1017, the file formatter
1020, and the customer web interface 1009.
[0232] The update manager 1006 selects the appropriate
transmissions medium, controls the packaging of data and schedule
delivery of the data (via the selected transmission medium) to the
wireless transmission apparatus 1010 and/or the Remote Relay Unit
(RRU) 1900, for subsequent delivery to the device 100. The update
manager includes an update scheduler 1025, an input interface 1026,
and an erasure correction routine 1027, a transmission scheduler
1028, in-field monitors 1029, and a re-submitter and diagnostics
system 1030, in addition to a transmission copies database 1031.
The update scheduler 1025 is a software process that runs
continuously reading the master database 1007 to retrieve the
schedule for file transmission and sending requests to the input
interface 1026 as required. The input interface 1026 is an
executable program that extracts prioritised and formatted data
from the current transmittable file storage database 1021. This
data is passed directly through to the erasure correction routing
1027. As data is passed through this routine, it is prefixed with
the appropriate device header information.
[0233] The erasure correction routine 1027 transforms the input
data into a number of redundant blocks (or error correction
fragments) that improve the reliability of reconstruction of the
original data, provided a given number of the redundant blocks are
received unchanged by the device 100. The redundant blocks may be
generated using the process described above in relation to FIGS. 11
and 12. The redundant blocks are packaged and sent to the
transmission scheduler 1028 individually. The transmission
scheduler 1028 selects the appropriate medium for transmission
(e.g. a wireless network used by the wireless transmission
apparatus 1010). The transmission scheduler 1028 receives and
stores packets passed from the erasure correction routine 1027, and
periodically forwards the highest priority data packets to, for
example, the wireless transmission apparatus 1010 for transmission
via the selected transmission medium (e.g. a wireless network). A
copy of the packets sent to the wireless transmission apparatus
1010 is written into the transmission copies database 1031 and
removed from storage. The in-field monitors 1029 are modified
versions of the client device 100, with modems attached to a
dedicated link to, for example, the PSTN network. These monitors
receive all data transmissions to the devices 100 in their
locality. This data is forwarded to the resubmitted and diagnostics
system 1030 via the PSTN network.
[0234] The resubmitter and diagnostics system 1030 includes
multiple input lines from the PSTN network 1004, and executes a
comparison process per input line and a single resubmit process.
The comparison process for each, input line compares data received
by the corresponding in-field monitor 1029 with data written to the
transmission copies database by the transmission scheduler 1028.
The resubmit process performed by the resubmitter and diagnostics
system 1030 periodically reviews the comparison results and
resubmits to the input interface 1026 any packets or files as
required. The in-field monitors 1029 also transmit information for
the purpose of assessing the performance of the selected
transmission medium/network (e.g. by, transmitting data via
antennas 1098). All information of this nature is stored in the
master database 1007 for analysis. The information is used on an
ongoing basis to ensure that the selected transmission
medium/network performs at a required specification.
[0235] The file loader 1019 and configuration manager 1018 maintain
the information stored in the master database 1007. The file loader
1019 performs routine loading of data contained in files with known
formats. The configuration manager 1018 is used for manual entry of
data that is not available in files, and for manual adjustment of
other data as required via an internal database maintenance module
1032.
[0236] The performance monitor 1008 is a group of processors that
monitor the operation of the support systems on a day to day basis.
The performance monitor 1008 provides a means of visual inspection
of resource usage levels, as well as standard reports at periodic
intervals. It also issues warnings when the system becomes
overloaded in any way. Components that are monitored in this way
include: [0251] (1) Files transmission. The number of errors and
hence retransmissions are monitored to determine transmission
quality in the various areas covered by the network. [0252] (2)
Database host. System resources on the database hose machine(s)
such as CPU, disk and memory are monitored to ensure performance
targets are met. [0253] (3) Database server. Inbuilt capabilities
of the database service software are used to ensure that the
database is running efficiently. [0254] (4) Web host. System
resources on the web-server host machine(s) such as CPU, disk and
memory are monitored to ensure performance targets are met. [0255]
(5) Web server. The inbuilt capabilities of web server software are
used to ensure that the database is running efficiently.
[0237] The wireless transmission apparatus 1010 uses the air
protocol and enables information to be efficiently distributed from
a central location to a large number of devices 100 using, for
example, the existing pager communication network.
2.4 The Remote Relay Unit (RRU)
[0238] The remote relay unit 1900, as shown in FIG. 22, is a
handheld device that can be used by service personnel or a user of
the device 100 to transmit file fragments to the device 100. The
RRU 1900 is particularly useful in situations where the wireless
communications network, eg the pager network, is unable to
communicate with the device 100, for example if coverage of the
network does not extend to the location of the device 100. The RRU
1900 is able to obtain the file fragments from the update manager
1006 using online connection methods, or by direct connection, to
the transmission system 1910 (e.g. via the data network 1099 as
shown in FIG. 23). Once the RRU has received the file fragments,
the fragments can be sent to the device 100 using a short range
wireless communications protocol, such as Bluetooth or ZigBee
(http://www.zigbee.org), or an infrared protocol, such as one that
complies with the standards of the Infrared Data Association (IrDA)
(http://www.irda.org). One RRU 1900 may be used to communicate and
configure a number of devices 100. The RRU 1900 may also be used to
program and/or control other items in a user's household, such a
VCR, DVD, TV, lights etc, either autonomously or by user
intervention. The control of these items, together with the devices
100, can be based on data dynamically available and updated from
the central transmission system 1910. The data and file fragments
can be delivered to the RRU 1900 based on user preferences stored
as part of the transmission system, or based on direct instructions
from the user, eg by interrogating the transmission system over the
Internet.
[0239] The RRU 1900 is also configured to enable information to be
extracted from local units, such as the household items and the
devices 100, so as to obtain diagnostic status, statistic and event
log data. A wireless interface, such as an IrDA or ZigBee
interface, can be used for retrieval of the information, but again
may be under the direct control of the user or performed
autonomously when a connection can be established. Information
retrieved from the local unit is stored on the RRU 1900 and then
made available to be sent to the central transmission system 1910
or for upload to a user's computer device, e.g. personal computer
system or PDA. The computer device is able to access the RRU 1900
using an IP connection.
[0240] The RRU 1900 includes a communications module 1902 that
enables the RRU 1900 to connect to the transmission system using at
least one of the following: GPRS; CMDA-1X; Ethernet; WLAN; or IP
Dial-up. The connection normally would be always online. If not,
the connection is either established by either the central server
1005 or the RRU 1900 on demand. The central server 1005 would
typically use the IP based protocols for connection, but a simplex
download connection to the RRU 1900 would suffice.
[0241] The RRU 1900 includes a flash file system 1904 that allows
information to be received locally in addition to being downloaded
from the central server transmission system 1005. The file system
1904 is scaled appropriately to cater for the number and size of
files required. The size of this file system 1904 may directly
relate to the number of local units 100 associated with it.
[0242] The RRU 1900 has a second communications module 1903 to
allow information to be distributed to its associated local units,
ie the devices 100, and support retrieval of information from the
local units. The second module 1903 provides the wireless
communications interfaces, such as ZigBee, IrDA, and/or pager
(POCSAG) interface.
[0243] The RRU 1900 has a microcontroller system 1905 that
maintains a real-time clock for synchronising the time in the
associated local units when required. The clock is set using a
network time protocol from the server transmission system 1005. The
system 1905 includes a user display that can be used to show status
information and to facilitate the user interface. The system 1905
also has buttons and/or a keypad to provide user
interaction/control.
[0244] A RRU 1900 can be associated with a number of user local
units that include third-party consumer devices (e.g. VCRs, TVs) in
addition to the client processor devices 100. The user registers
the make and model number of each of the consumer devices on the
central server transmission system 1005 via the web interface 1009.
Depending on support for the third party devices, configuration
information is generated on the central server 1005 and sent to the
RRU 1900 to allow the devices to be controlled. The serial number
of each of the client processor devices 100 is similarly registered
on the web interface 1009. The central server 1005 controls the
number and type of associated devices 100 depending on the
resources (power, memory and communications bandwidth) associated
with the remote relay unit 1900. A particular device 100 is
normally associated with a single remote relay unit 1900.
[0245] In order that a relay unit 1900 can communicate with a
particular device 100 it may, depending on the transmission method,
need to be in the vicinity or close line-of-sight of the device.
This means the RRU 1900 is normally a portable device similar to a
mobile phone. It is returned to a cradle 1906 to recharge and/or
make a wired connection to server transmission system 1005, if
required.
[0246] In order that the client processor devices 100 can
communicate with the relay unit 1900, they are provided with the
I/O module 130. This provides a corresponding communications
interface 1908, such as ZigBee or IrDA. These interfaces support
relatively high-speed bidirectional communication and this ensures
that the data is transferred in a reliable and timely manner. The
device 100 monitors the alternate interface 1908 and communications
over this interface 1908 would normally override that received on
the pager receiver 110.
[0247] After the RRU 1900 and associated local units are
registered, the file system 1904 is then able to accept updates
from the central server 1005. The user may chose to make changes
(i.e. configuration or otherwise) using the web interface 1009 or
new data may be available from the server 1005 depending on the
service(s) the user is subscribed to. Subscription data indicating
the services subscribed by a user may be maintained, for example,
on the central server 1005. If new data is available the central
transmission server 1005 checks the online status of user's the RRU
1900 and indicates this status to the user during the time the user
is using the web interface. Once the changes are sufficiently
completed the server 1005 attempts to download the data to the
remote unit 1900. Once the data is fully downloaded, the RRU 1900
indicates this as fresh data available (e.g. as a flashing light or
textual display) and then proceeds to automatically transfer the
data to the associated local units. If the RRU 1900 cannot be
contacted from the server 1005, the user needs to take action such
as manually starting the connection by returning the RRU 1900 to
the cradle 1906. If the RRU 1900 has data for an associated client
processor device 100 but cannot connect to it, the user is
instructed to take appropriate action (i.e. to bring the remote
relay unit 1900 to the client processor device 100). The client
processor device 100 may similarly indicate that it has received
fresh data (e.g. by flashing a light).
3. Telecommunications Signal Carrier Selection
[0248] The client processor device 100, when used as part of a call
management system 1001 acts as a customer interface device 1002, as
shown in FIG. 24, for connection between a telephone or other local
telecommunications device 1003 and a telecommunications network
1004, such as a Public Switched Telephony Network (PSTN). The call
management system includes the server transmission system having a
file collator 1005, an update manager 1006, a master database 1007,
performance monitor 1008 and a wireless transmission apparatus
1010. The server transmission system also provides the customer web
interface 1009.
[0249] The client device 1002 is the physical and logical component
of the call management system that enables a user of local
telecommunications device 1003 to gain real time access to the most
cost effective telecommunications signal carrier available at any
given time. The device 1002 has a line interface 1016 that operates
as an I/O unit 130 and is connected between the telecommunications
device 1003 and the network 1004, as shown in FIG. 24.
[0250] As shown in FIG. 24, the line interface 1016 of the client
processor device 1002 communicates with the processor core 102 of
the microcontroller 150 on a data and address bus 114 and a control
bus 2221. The line interface 1016 receives control signals
generated by the processor core 102 from the control bus 2221,
which triggers the line interface 1016 to send or receive data via
the data and address bus 114. The client processor device 1002 also
includes an alpha numeric display 1033 for the user to view current
information available, and the display 1033 communicates with
processor core 102 via the data and address bus 114.
[0251] The virtual machine VM of the client device 1002 executes an
application (e.g. 2330 as shown in FIG. 25) transmitted from the
central server 1005 to perform the selection of an optimal
telecommunications signal carrier by controlling the line interface
1016 at the time when a call is desired to be placed from the local
telecommunications device 1003. For example, the VM restarts the
application whenever the telecommunications device is detected as
going "off hook".
[0252] The line interface 1016 comprises electronic circuitry
interposed between the local telecommunications device 1003 (such
as a telephone, facsimile machine, modem etc) and a telephony
exchange, PABX system or other network mode connecting the local
telecommunications device 1003 to the telecommunications network
1004. The line interface 1016 detects the "of hook" or line looping
state, either of the local telecommunications device 1003 or of
other equipment if the phone line is in use elsewhere. The line
interface 1016 detects and decodes DTMF digits dialed by a user
from the local telecommunications device 1003. The line interface
device 1016 additionally isolates the audio path on the telephone
line from the local telecommunications device 1003 when instructed
to by the virtual machine, so as to be able to intercept and
prevent dialed digits from reaching the telephone exchange, PABX or
other network node that would ordinarily receive the call. The line
interface device 1016 additionally dials DTMF digits when
instructed to do so by the virtual machine executing in the
processor core 102, in order to make calls autonomously or whilst
the local telecommunications device 1003 is off-hook and isolated
from the telephony exchange or PABX.
[0253] The master database 1007 is used to record all structure
information used to control operation of the call management
system.
[0254] Typically, the following information is stored in the
database 1007. [0255] (1) Call rates for each carrier. [0256] (2)
Carrier details. [0257] (3) Exchange codes. [0258] (4) Customer
details and preferences. [0259] (5) Carrier rate structure (for
example, peak and off peak rates) [0260] (6) Field monitor
locations and connection details. [0261] (7) File transmission
states [0262] (8) File transmission schedules. [0263] (9) Customer
billing data. [0264] (10) Carrier billing data.
[0265] While the device 1002 shown in FIG. 23 enables the
interconnection of the local telecommunications device 1003 with a
PSTN network 1004, the client processor device 1002 may further
enable connection of the local telecommunications device 1003 to a
remote telecommunications device via a selected telecommunications
signal carrier within a same telecommunications network (such as
the PSTN network 1004), or across two or more different
telecommunications networks (such as a mobile, ISDN, frame,
Internet, satellite, voice-over-ip, or internal network).
Accordingly, the carrier identifiers, variable costing information
and other data maintained in the memory 120 of the client processor
device 1002 (as updated from the central server 1005) may relate to
a single telecommunications network, or to different
telecommunications signal carriers for different types of
telecommunications networks (i.e. telecommunications networks that
conform to differing protocols). Both types of client processor
devices have the same hardware and logical architectures as shown
in FIGS. 24 and 25, respectively. Additional tasks are included for
sue by the application files 2330, such as a line task 2321 for
controlling the line interface 1016 to connect calls over one or
more different telecommunications networks.
[0266] In operation, the device 1002 performs the following basic
steps whenever a call is attempted to be established from the local
telecommunications device 1003 to a remote telecommunications
device: [0267] (1) Determine any parameters supplied by the user
for the call, such as the maximum/minimum length of call or call
cost. These parameters come from the user at the time the call is
made, for example, by entering a special key sequence. [0268] (2)
Determine the customer's general parameters such as call blocking,
average call time or preselected carrier. These parameters are
stored in the information storage system. [0269] (3) Read any key
pressed by the user until the nature of the call, for example
local, neighborhood, national, international, etc., can be
determined. [0270] (4) Determine the destination of the call--this
process may differ depending upon the nature of the call cost for
each carrier available to the user. [0271] (5) Determine the
cheapest carrier for use in placing the call. [0272] (6) Send an
override code which will select the cheapest carrier to the line
interface 1016, then send the units dialed so far and place the
device 1002 in a stage which will send all further digits straight
through to the line interface 1016 so the number can be completed
(if it has not already).
[0273] The above described operation relies upon detecting the
digits or other identifiers entered by a user at the local
telecommunications device 1002, and the comparison of these digits
(as they are dialed in real time) with data maintained in the
information storage system 120. The data maintained in the
information storage system 120 indicates the identity of various
telecommunications signal carriers, the variable costing
information for each carrier, and the location of telephone
exchanges or other billing calculation nodes within one or more
telecommunications networks.
[0274] FIG. 26 is a flow diagram providing a high level overview of
the process 7600 performed by the device 1002 for selecting an
optimal telecommunications signal carrier. Process 7600 begins at
step 1060, where a connection is established between the local
telecommunications device 1003 and the PSTN 1004 or other
telecommunications network to which the local device 1003 is to be
connected, and the use has not yet entered a remote
telecommunications device identifier (such as dialed numerical
string of digits). At step 1061, a remote telecommunications device
identifier is entered from the local telecommunications device 1003
by a user. As each digit is entered, an attempt is made to resolve
the sequence of digits thus far entered until a corresponding
telephone exchange or other network billing node is able to be
identified at step 1062. If it is determined at step 1063 that
resolution of the entered string of digits was possible and a
billing calculation node is able to be identified, then target data
for that node is retrieved from the memory system 120 in step 1064.
The telecommunications signal carrier providing the cheapest call
cost is then determined in step 1065. At step 1066, a determination
is made as to whether the device 1002 is adapted to route the call
from the local telecommunications device 1003 via multiple types of
networks (FIG. 32). If this is the case, each of these interface
options are examined at step 1067 to determine whether a cheaper
call may be obtained via these interface options
[0275] If the interface options have not been installed, a carrier
identifier string corresponding to the selected optimal
telecommunications signal carrier is then transmitted to a
telephone exchange associated with that carrier, together with the
string of digits already composed by the user at the local
telecommunications device, at step 1068. At step 1069, further
digits required to identify the remote telecommunications device
with which a connection is to be established (e.g. an area code or
country code) are transmitted directly to the telephone exchange
associated with the selected carrier. The process 7600 is
terminated at step 1070.
[0276] If it was determined at step 1063 that resolution of the
string of digits entered at the local telecommunications device by
a user to identify a telephone exchange or other billing
calculation node is not possible, the digits received are
transmitted at step 1064a the telephone exchange of the default
telecommunications carrier, and then step 1069 is carried out.
[0277] FIG. 20 shows a flow diagram of the process 7700 for
examining the application files stored in the memory 120 of the
client processor device (e.g. 1002) for validity prior to
execution. Process 7700 begins at step 1071. Once the device 1002
is initiated, step 1071 proceeds to step 1071a, where the next
application file in the init-tab file maintained in the memory
system 120 is read and a determination is made as to whether that
application file is valid. If the application file is not valid,
process 7700 finishes at step 1074. If the application files are
valid, then at step 1072 the VM determines using the init-tab file
whether all data files required for execution of that latest
application file are valid and complete. If this is the case, then
step 1072 proceeds to step 1073 to execute the current application,
and then process 7700 finished at step 1074. However, if step 1072
determines that the set of data files is not complete or valid for
a particular current application file, then step 1072 proceeds to
step 1071a, where the next most recent application file is examined
and a determination made as to whether the data files required for
execution of the next version of the application file are current
and valid.
[0278] In the case of a PSTN network, call costing is determined on
the basis of the geographical location of the telephony exchange to
which the local telecommunications device 1003 is directly
connected. In Australia, customer telephone numbers are grouped
into exchange service areas, which are in turn grouped into zones.
Zones are grouped to form charging districts and call costs are
calculated based on the distance between the zone and/or district
charging point of the A-party number and the B-party number. The
chargeable distance for a call is based on the distance between the
zones or districts the parties are in, rather than the actual
distance between the callers. That is, each zone or district has
just one centrally located geographically point for measurement of
the distance to other zones or districts. Accordingly, step 1062 in
FIG. 26 requires the resolution of the digits entered by the user
of the local telecommunications device 1003 in order to enable
identification of the telephony exchange to which the local
telecommunications device 1003 is directly connected. A
determination can then be made of the district or zone within which
that telephony exchange is located in order to derive a call
distance upon which a call charge may be calculated. As each digit
is dialed by the user, an attempt is made to resolve the string of
digits already entered so that the resolution to identify the
telephone exchange in question, and hence determine the
corresponding billing calculation location, without requiring entry
of the entire string of digits identifying the remote
telecommunications device.
[0279] In Australia, there are some 4,500 telephony exchanges,
grouped into various precincts, zones and districts. There are only
a small number of precincts (currently 118) but over 2,000 zones in
which the telephony exchanges may be located. Each telephony
exchange is identified by an exchange code. The exchange code forms
part of the telephone number identifying the telecommunications
devices connected to the network. For example, in Australia the
telephone number +61-03-9851-0000 includes an international access
code (represented here by the symbol "+"), the country code 61 for
Australia, the state (or area) code of (0)3 for Victoria, the
exchange code 9851 and a terminal code 0000 to identify the
communications device connected to each exchange. An exchange code
file is maintained in the information storage system 120. The
exchange code file contains a series of encoded exchange codes
along with the identification of the exchange for each code. The
exchange code resolution process determines whether a given string
of digits identifies a specific exchange and if so, finds the
identifying details for that exchange.
[0280] The exchange code file is not a simple list of exchange
codes, but includes several space optimizations. The exchange code
file takes advantage of the fact that assorted series of codes can
be more efficiently expressed as a series of differences to the
preceding code. Each code entry in the exchange code file therefore
consists of a number of characters to retain from the previous
code, and the new characters to add to that code. Ranges of codes
are compressed into a single entry since all codes consist of
decimal digits only, pairs of digits can always be expressed as a
single byte using four bits for each digit.
[0281] For example, the series of codes 0234, 0235, 02351, 02353,
02354, 02355, 02366 and 02367 if stored in an uncompressed form,
would require 4+4+5+5+5+5+5+5=38 bytes. If differences only are
considered, the following stored codes can be used: [0282] 0, 0234
Keep none, add 4=0234 [0283] 3,5 Keep 3 (023), add 5=0235 [0284]
4,1 Keep 4 (0235), add 1=02351 [0285] 4,3 Keep 4 (0235), add
3=02353 [0286] 4,4 Keep 4 (0235), add 4=02354 [0287] 4,5 Keep 4
(0235), add 5=02355 [0288] 3,66 Keep 3 (023), add 66=02366 [0289]
4,7 Keep 4 (0236), add 7=02367 This requires only 5+(6*2)+3
bytes=20.
[0290] Considering ranges, where a range is a series of codes with
successive values that belong to the same exchange and have the
same length, gives: [0291] 0.0234- Keep none, add 0234, run the
last digit to 5=0235 [0292] 4,1-1 Keep 4 (0235), add 1=02351, no
range [0293] 4,3-5 Keep 4 (0235), add 3=02353, run the last digit
to 5=02354, 02355 [0294] 3,66-7 Keep 3 (023), add 66=02366, run the
last digit to 7=02367 This requires only 16 bytes
[0295] The space can be halved again by considering the above
sequence in 2 character pairs: [0296] 00 23 45 41 14 35 36 67
[0297] These pairs can be represented as hexadecimal byte values,
eg 00=0, 23=35, 45=69, etc., and one byte written for each
pair.
[0298] The data section of the exchange file is essentially made up
of a sequence of blocks with the same internal structure. That
structure is as follows:
[0299] Code Block: <Keep><New Digits><Range
Digit><Exchange Type><Exchange ID>, where
[0300] <Keep>: Number of digits to keep from the previous
entry, first entry in file must have Keep=0.
[0301] <New Digits>: Sequence of digits to add to those kept
from the previous entry.
[0302] <Range Digit>: If non-zero, signifies that entry
should match all the codes between the code derived from Kept+New
and Kept+Range_End, where Range_End is New with the last digit
replaced by the Range Digit
[0303] <Exchange Type>: Indicates whether this exchange is a
Zone, Precinct or Other (eg ISD, emergency, mobile, etc.).
[0304] <Exchange ID>: The index of the exchange within its
type; if the Type is Zone, the Exchange ID will be a Zone
Index.
[0305] An example of the exchange code file is shown in Table
2:
TABLE-US-00019 TABLE 2 Exchange Code Type ID 000 Other 1 0011 Other
2 03980 Zone 1 039802 Zone 2 03985 Precinct 1 03986 Precinct 1
03987 Precinct 1 0399 Zone 3 03992 Zone 4 03993 Zone 4 0419 Other 3
131 Other 4
[0306] In order to simplify the example, the final list of codes is
not converted to hexadecimal byte values. The Exchange Type values
used is 1 for Other, 2 for Zone and 3 for Precinct. The data set
shown in Table 2 is used to generate the following data string:
[0307] 0000011211012139800215201245731390234232414190130131014
[0308] Splitting this string into code blocks: [0309] 0000011211012
13980021 52012 45731 39023 42324 1419013 0131014
[0310] Tagging each part of each block with K for Keep, N for New
digits, R for Range
[0311] Digit, T for Exchange Type and I for Exchange ID: [0312]
0000011 211012 13980021 52012 45731 39023 42324 1419013 0131014
KNNNRTI KNNRTI KNNNNRTI KNRTI KNRTI KNRTI KNRTI KNNNRTI KNNNRTI
[0313] The data is processed from the beginning from a block with a
Keep value of 0, since all other blocks depend on the preceding
blocks back to the most recent Keep=0 block.
[0314] The device 1002 uses the exchange code file to determine the
exchange code applicable to a string of digits (the "Target")
entered by the user. Each code block is read in turn, the complete
exchange code derived for that entry by combining the previous code
with the Keep and New Digits values for the new block. The target
string is then compared to the file exchange code, possibly
including additional file codes as dictated by the Range Digit.
There are four possible outcomes from this comparison: [0315] (1)
Continue; the point in the file has not been reached where the
target string should occur, and the next block. [0316] (2)
Incomplete; read past the point where the target string should have
occurred, but adding more digits to the target string and
re-starting the search may result in an exact match. [0317] (3)
Valid; found an exact match and should use the exchange data for
the current block [0318] (4) Invalid; read past the point where the
target string should have occurred, and the target string cannot be
made to match by adding to digits to it.
[0319] In practice, the client processor device 1002 attempts to
resolve the exchange after each DTMF digit is received. Each
attempt will produce one of outcomes 2, 3 and 4 above; code
resolution will continue to take place until one of outcomes 3 and
4 arise.
[0320] For an example, consider the phone number (03) 9851 0000. As
this number is dialed, resolution will be attempted after each new
digit, using all the current digits each time. The results from
these attempts are shown in Table 3 below:
TABLE-US-00020 TABLE 3 Last block Last exchange New digit Target
string Outcome read code 0 0 Incomplete 000011 000 3 03 Incomplete
13980021 03980 9 039 Incomplete 13980021 03980 8 0398 Incomplete
13980021 03980 5 03985 Valid 45731 03985-7
[0321] The pattern of a successful resolution is shown above.
Repeated "Incomplete" outcomes, where the target string is a subset
of the exchange code of the last block read. The outcome of the
comparison between the target and all the codes preceding the final
one would have been "Continue", hence the next block was read. An
exit is made after these final blocks, since the file exchange code
is greater than the target code, but the potential for a match is
always there (since the target string and the last exchange code do
match over the length of the target), and does eventually occur
(outcome "Valid").
[0322] For a second example, consider the outcomes for phone number
0398810000.
TABLE-US-00021 TABLE 4 Last block Last exchange New digit Target
string Outcome read code 0 0 Incomplete 000011 000 3 03 Incomplete
13980021 03980 9 039 Incomplete 13980021 03980 8 0398 Incomplete
13980021 03980 8 03988 Invalid 39023 0399
[0323] As shown in Table 4, the outcomes for the first 4 digits are
the same, as they would be. After the fifth digit is entered, the
final block comparison returns "Invalid", as there is no way a
match for a target starting with "03988" will occur if one has not
been found before file code "0399", regardless of what subsequent
digits may be entered.
[0324] The application automatically adds the local state code (eg
03 for Victoria) to numbers that do not already begin with 0 or 1
before attempting exchange code resolution.
[0325] FIG. 28 shows a flow diagram of the process 7800 for
resolving the exchange code as described above. Process 7800
processes a series of digits as currently entered by the user
(defined as the "target" string). Each time a user enters a new
digit, the target string is updated to include the new digit, and
process 7800 processes the updated target string again starting
from step 1080. Process 7800 begins at step 1080 by accessing the
target string of digits entered by the user. Step 1080 proceeds to
step 1081 to attempt to retrieve an entry from the exchange code
file, the entry corresponding to a block of code (as described
above) that represents an exchange code. If an entry cannot be
found in the exchange code file at step 1081, process 7800 ends by
determining that the target string is invalid. Otherwise, if an
entry is found, step 1081 proceeds to step 1082.
[0326] Step 1082 generates an exchange code based on the code for
the current entry retrieved at step 1081, and determines whether
the exchange code for the current entry is past the point where the
target string should have occurred. As shown in Table 2, the
entries in the exchange code file are arranged in descending order
based on the exchange code value for each entry. Step 1082
determines that the exchange code for the current entry is past the
point where the target string should have been if the exchange code
for the current entry includes other digits in addition to all the
digits in the target string. For example, where a target string
includes only the digits "03", a current entry having a
corresponding exchange code of "03980" is past the point where the
target string should have occurred in the descending order of
exchange code values in the exchange code file. If the exchange
code for the current entry has not passed the point where the
target string should have occurred, step 1802 proceeds to step 1803
to determine whether the digits in the target string exactly
matches the exchange code for the current entry. If so, the data
characterising the exchange identified as corresponding to the
current entry is retrieved at step 1084 (e.g. by using process 7900
as shown in FIG. 28), and process 7800 finishes by determining that
the target string are valid.
[0327] However, if there is no match at step 1083, step 1083
proceeds to step 1081 to attempt to retrieve the next entry in the
exchange code file as the new current entry, which corresponds to
another exchange code. If the next code is unable to be read at
step 1081, process 7800 ends and the series of digits entered by
the user is determined to be invalid. If it is determined at step
1082 that the exchange code corresponding to the current entry has
past the point where the target string should have been in the
exchange code file, determination is then made at step 1085 as to
whether additional digits in the target string would assist in
resolving the code. For example, step 1085 assesses whether the
exchange code for the current entry includes additional digits
which do not present exist in the target string. If so, step 1085
determines that additional digits in the target string might be
helpful to resolve the exchange code, and process 7800 ends by
determining that the digits entered by the user is incomplete, and
process 7800 starts again on receiving an updated target string
that includes a new digit. However, if step 1085 determines that
this is not the case, process 7800 ends and the digits entered by
the user are determined to be invalid.
[0328] Once the exchange code corresponding to the exchange to
which the remote telecommunications device is connected has been
thus identified (e.g. at step 1064 in process 7600 and step 1084 in
process 7800), data corresponding to that exchange is retrieved by
the virtual machine from the memory system 120. An exemplary file
for an exchange generated from the retrieved data may be as
follows: [0329] Zones: [0330] Aldersyde [0331] Exchange codes:
089600(0-2), 089642(6-7) [0332] Adjoining zones: Brookton,
Bullaring, Pingelly, Pingelly East, South Quairading, Yealering
[0333] District: York [0334] Charge Point: 32<B0>25'30''S
117<B0>18'30''E [0335] Community access: Narrogin [0336]
Alectown [0337] xchange codes: 0268005, 0268613, 026865 [0338]
Adjoining zones: Baldry, Bindogundra, Bogan Gate, Bruie Plains,
Parkes, Peak Hill [0339] District: Parkes [0340] Charge Point:
32<B0>56'30''S 148<B0>13'0''E
[0341] From this file, the charge point from which the chargeable
distance for the call will be calculated is identified.
[0342] The exchange code to which the local telecommunications
device 1003 is directly connected is identified in an identical
manner to the process described above, and the charging point of
the zone, district, or precinct within which that telephony
exchange is located similarly determined. Comparison of the
charging points in the zones, districts within which the telephony
exchanges to which the local and remote telecommunications devices
are directly connected are then used to determine the chargeable
distance for the call.
[0343] The manner in which the file loader 1019 extracts data from
the exchange structure data 1023 in order to derive the one or more
files, each categorising a different telephony exchange, is
illustrated by process 7900 as shown in FIG. 28. The file loader
1019 is responsible for extracting data from one or more files
(e.g. flat text files or HTML files) provided by telecommunications
infrastructure providers, and inserting the extracted data into the
database 1007. Examples of information contained in these files
includes exchange data, carrier call rates and lists of current
international dialing codes. Carrier information 1024 is also
extracted in this same manner.
[0344] Process 7900 begins at step 1090, where the file loader
process 1019 is initiated by receiving a file from a
telecommunications infrastructure provider. At step 1091, a first
line of data is extracted from the relevant file from the
telecommunications provider. If it is determined at step 1092 that
the data being examined is district data, then that data is
formatted at step 1093 into district data. Similarly, if it is
determined at step 1094 that the data characterises the zones
adjoining the zone or other region under consideration, that data
is extracted and formatted at step 1095 as zone data. A similar
process is followed at steps 1096 to 1101 for charge point data,
community access data and any other types of data in the file.
Steps 1093, 1095, 1097, 1099 and 1101 all proceed to step 1091 to
select the next line of data from the file. When step 1102
determines that the end of file has been reached, the formatted
data is loaded into the database at step 1103, and the file loader
process 7900 finishes at step 1104. Otherwise, if step 1102
determines that the end of the file has not been reached, step 1102
proceeds to step 1091 to select the next line of data from the
file
[0345] Once the chargeable distance for the call in question has
been determined, the carrier providing the cheapest call cost for
that chargeable distance is then selected. FIG. 30 is a flow
diagram of the carrier selection process 8000 for achieving this.
Process 8000 begins at step 1121 by determining the current date
and time (e.g. by querying the clock task 225 of the client
processor 1002 for the current time). Step 1121 proceeds to step
1122 to retrieve data for a particular telecommunications carrier
(e.g. the data generated by process 7900 that is stored in the
memory 120 of the client device 1002 by way of update data
transmissions from the update manager 1006). For example, step 1122
retrieves the call tariffs at different time bands (e.g. peak and
off peak), the distance segment, and the rate for the time band and
distance.
[0346] At step 1123, the current time determined at step 1121 is
used to search the data retrieved at step 1122 for a particular
carrier to determine the call rates applicable for the current time
band. Then, at step 1124, the distance segment of the call for the
carrier is determined (e.g. based on the charge points as described
above). Step 1125 then determines the carrier's rate based on the
time band and distance determined at steps 1123 and 1124
respectively. The rate determined at step 1125 is used to generate
the call cost at step 1126. Step 1127 stores the call cost
generated at step 1126 for each carrier, and compares the call cost
generated for the carrier currently being analysed with the stored
call cost data. Based on this comparison, step 1127 determines
whether the call cost generated for the carrier currently under
analysis is the best (i.e. the lowest) cost determined so far. If
this is the best cost yet, as determined at step 1127, the carrier
is recorded at step 1128 as being the best or optimal carrier. If
this is not the case, step 1127 proceeds to step 1122 to retrieve
data for another carrier, and process 8000 repeats again until the
data for all the carriers (as stored in the database 1007 and
memory 120) have been processed. Once no more carriers are
available for processing, any digits subsequently dialed by the
user are passed directly through to the telephony exchange
associated with that optimal carrier at step 1129. The carrier
selection process is terminated at step 1130.
[0347] FIG. 30 illustrates the steps of the process 8100 performed
by the client processor device 1002 when an incoming call is
received, and redirected to a remote telecommunications device 1050
(as shown in FIG. 23) associated with the owner of the local
telecommunications device 1003. In this example, the device 1050
enables connection of the local telecommunications device 1003 to
the remote telecommunications device 1050 via multiple
telecommunications networks, such as a mobile telecommunications
network or a Voice-Over-IP (VoIP) network. Process 8100 starts at
step 1140, where an incoming call is received at the local
telecommunications device 1003. The client processor device 1002
then acts to redirect that incoming call to a remote
telecommunications device 1050 associated with the owner of the
local telecommunications device 1003. Initially, at step 1141, the
customer interface device 1050 determines the called party (or
B-party) number to which the incoming call has to be redirected. At
step 1142, the digits forming part of the called party (or B-party)
number are analyzed, in the manner described in relation to FIG.
28, in order to resolve the telecommunications exchange or other
telephony node to which the call is to be routed.
[0348] At step 1143, a determination is firstly made as to whether
the device 1002 includes interfaces to multiple telecommunications
networks. A determination is also made as to whether the B-party
number is supported by any of the telecommunications networks to
which the customer interface device is able to be connected. The
device 1002 also determines (in a manner similar to that described
in relation to FIG. 30) whether the various alternate interfaces
providing access to telecommunications networks other than the
conventional PSTN network 1004 will provide a cheaper call
establishment cost If the device is determined in step 1143 as not
supporting multiple telecommunications networks, a carrier
identifier string corresponding to the optimal selected
telecommunications signal carrier within the PSTN network 1004 is
then transmitted to telephone exchange associated with that
carrier, together a string of digits from the calling party
(A-party) number, at step 1144.
[0349] At step 1145, further digits required to identify the remote
telecommunications device to which the call is to be redirected are
transmitted directly to the telephone exchange associated with the
selected carrier. This process is terminated at step 1146.
[0350] However, if it is determined at step 1143 that multiple
interfaces are available, then a determination is made at step 1147
as to whether the selected optimal interface is associated with a
mobile network. If this is the case, a connection is established at
step 1148 to the mobile interface, then, at step 1149, the B-party
number is transmitted to the mobile network for call completion. At
step 1150, further digits required to identify the remote
telecommunications device to which the call is to be routed are
then transmitted directly to an exchange within the mobile network.
This process 8100 is terminated at step 1150.
[0351] If step 1147 determines that the interface associated with
the telecommunications network selected at step 1143 was not
associated with a mobile network, step 1147 proceeds to step 1152
to determine whether the interface associated with the
telecommunications network selected at step 1143 was associated
with another type of interface (e.g. a GPRS network). If so, step
1152 proceeds to step 1153, where the appropriate sequence of steps
are performed by the device 1002, depending upon the
telecommunications network selected. For example, this may involve
performing steps similar to steps 1148 and 1149 by connecting to
the other chosen interface (e.g. a GPRS interface) and sending
identification information to that interface for the purpose of
establishing a call. However, if step 1152 determines that a
Voice-over-IP (VoIP) interface has been selected at step 1143, then
step 1152 immediately proceeds to step 1154.
[0352] At step 1154 it has been determined the call is to be routed
via a VoIP network. The A-party number is resolved into a network
address at step 1155. This network address is used at step 1156 to
establish a connection with a computing device or other network
entity within the VoIP network, and the incoming call then is
redirected to that computing device.
[0353] FIG. 31 is a flow diagram of the process 8200 for handling
an alternative scenario to that shown in FIG. 30, in which a
telephone call is initiated from the local telecommunications
device 1003 and may be routed via one of multiple
telecommunications networks interfaced with the client processor
device 1002, rather than an incoming call being redirected by the
device 1002 as was previously described in relation to FIG. 30. In
FIG. 31, at step 1170, a remote telecommunications device
identifier string (such as a telephone number associated with a
remote telephone 1050 or other telecommunications device) is
composed at the local telecommunications device 1003 by a user. At
step 1171, a determination is made as to whether an interface to an
alternate network other than the PSTN network 1004 is available. If
no alternate telecommunications network is available, then at step
1172, a carrier identifier string corresponding to the selected
optimal telecommunications signal carrier within the PSTN network
1004 is then transmitted to a telephone exchange associated with
that carrier, together with the string of digits already composed
by the user at the local telecommunications device 1003. At step
1173, further digits required to identify the remote
telecommunications device with which a connection is to be
established are transmitted directly to the telephone exchange
associate with the selected carrier. The process 8200 is terminated
at step 1174.
[0354] At step 1171, a determination is also made as to whether the
alternate telecommunications networks are able to offer a cheaper
call connection cost, and whether the B-party number is supported
by any of the telecommunications networks to which the customer
interface device is able to be connected. At step 1175, if it is
determined that an acceptable alternate telecommunications network
through which the call is to be routed is a mobile network, then at
step 1176 a connection is established to the mobile interface, and,
at step 1177, the digits dialed by the user at the local
telecommunications device 1003 are transmitted through the mobile
interface. At step 1178, further digits that may be entered by a
user at the remote telecommunications device (for example in an
auto attendance application) are then transmitted directly to the
telephone exchange or other routing devices associated with the
selected carrier in the alternate telecommunications network. The
process 8200 is terminated at step 1179.
[0355] At step 1180, a determination is made as to whether the
selected alternate telecommunications network is neither a mobile
network or VoIP network (e.g. a GPRS network). If this is the case,
then, at step 1181, an appropriate sequence of operations are
performed by the device 1050 in order to direct the call through
that alternate network. Alternatively, if the alternate
telecommunications network is a VoIP network, as determined by step
1182, then the number dialed at the local telecommunications device
1003 is resolved into a network address at step 1183, and a
connection then established from the local telecommunications
device 1003 to a computing device or other network entity (for
example a VoIP telephony device) at that network address at step
1184.
4. Other Applications
[0356] Many other applications are available for the client
processor device 100 and the server transmission system. For
example, photocopier software can be upgraded remotely. Automotive
electronic system firmware can be upgraded without need to recall
the vehicle new features can be ordered for home appliances over
the Internet and delivered when the customer wants them. Electronic
billboards can be provided with new display information.
Supermarket price tags can be updated across the store or across
the country. RIFD readers can have new standards installed in the
field. Building and home systems can be switched off and on, such
as watering systems, alarm systems or lighting, using the devices
100
[0357] Many modifications will be apparent to those skilled in the
art without departing from the scope of the present invention as
herein described with reference to the accompanying drawings.
* * * * *
References