U.S. patent application number 15/423228 was filed with the patent office on 2017-11-30 for method and system for co-operative on-path and off-path caching policy for information centric networks.
This patent application is currently assigned to Tata Consultancy Services Limited. The applicant listed for this patent is Tata Consultancy Services Limited. Invention is credited to Bighnaraj PANIGRAHI, Hemant Kumar RATH, Anantha SIMHA.
Application Number | 20170346735 15/423228 |
Document ID | / |
Family ID | 57956160 |
Filed Date | 2017-11-30 |
United States Patent
Application |
20170346735 |
Kind Code |
A9 |
RATH; Hemant Kumar ; et
al. |
November 30, 2017 |
METHOD AND SYSTEM FOR CO-OPERATIVE ON-PATH AND OFF-PATH CACHING
POLICY FOR INFORMATION CENTRIC NETWORKS
Abstract
A method and a system is disclosed herein for co-operative
on-path and off-path caching policy for information centric
networks (ICN). In an embodiment, a computer implemented method and
system is provided for cooperative on-path and off-path caching
policy for information centric networks in which the edge routers
or on-path routers optimally store the requested ICN contents and
are supported by a strategically placed central off-path cache
router for additional level of caching. A heuristic mechanism has
also been provided to offload and to optimally store the contents
from the on-path routers to off-path central cache router. The
present scheme optimally stores the requested ICN contents either
in the on-path edge routers or in strategically located off-path
central cache router. The present scheme also ensures optimal
formulation resulting in reduced cache duplication, delay and
network usage.
Inventors: |
RATH; Hemant Kumar;
(Bangalore, IN) ; PANIGRAHI; Bighnaraj;
(Bangalore, IN) ; SIMHA; Anantha; (Bangalore,
IN) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Tata Consultancy Services Limited |
Mumbai |
|
IN |
|
|
Assignee: |
Tata Consultancy Services
Limited
Mumbai
IN
|
Prior
Publication: |
|
Document Identifier |
Publication Date |
|
US 20170237657 A1 |
August 17, 2017 |
|
|
Family ID: |
57956160 |
Appl. No.: |
15/423228 |
Filed: |
February 2, 2017 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
H04L 67/2842 20130101;
H04L 45/306 20130101; H04L 67/1097 20130101; H04L 45/38 20130101;
H04L 67/327 20130101 |
International
Class: |
H04L 12/747 20130101
H04L012/747; H04L 12/721 20130101 H04L012/721; H04L 12/927 20130101
H04L012/927; H04L 12/773 20130101 H04L012/773; H04L 29/08 20060101
H04L029/08; H04L 12/725 20130101 H04L012/725 |
Foreign Application Data
Date |
Code |
Application Number |
Jan 15, 2016 |
IN |
201621001593 |
Claims
1. A method to improve network performance by co-operative caching
content in a multi-router information centric network (ICN)
architecture, the method comprising: on-path caching, wherein a
plurality of edge routers of a multi-router network cache an
incoming content from a source router in the return path while the
content is being served to an user of the multi-router network; and
off-path caching, wherein a predefined central router of the
multi-router network caches one or more un-cached contents and one
or more filtered out staled contents of each of the plurality of
edge routers.
2. The method claimed in claim 1, further comprising: the plurality
of edge routers of the multi-router network are assisted by the
pre-defined central router of the multi-router network; and the
incoming content cached at atleast one of the plurality of edge
routers of the multi-router network is based on availability of
cache of the edge router.
3. The method claimed in claim 1, further comprising: computing
score of the incoming content to cache at each of the plurality of
edge routers of the multi-router network, wherein the computed
score is used to tag cache lifetime of the content; and replacing
stale cached content of the atleast one of the plurality of edge
routers, wherein the replacement is based on the lower cache
lifetime of the staled cache and incoming content.
4. The method claimed in claim 3, wherein the computed content
score is based on normalized distance between the source router and
serving router of the multi-router network and normalized frequency
of access of the content.
5. The method claimed in claim 1, wherein the caching of the one or
more filtered out staled contents from the plurality of routers and
the one or more un-cached contents is based on state information of
the plurality of edge routers, cache hit probability of the content
and average delay being experienced while submitting a content for
caching.
6. The method claimed in claim 5, wherein the state information of
the plurality of edge routers include arrival rate and diversity
index of the requests for content.
7. The method claimed in claim 5, wherein the caching of the one or
more filtered out staled contents from a plurality of routers is
based on a cache router score of each of the plurality of
routers.
8. The method claimed in claim 7, wherein the cache router score
comprising of scaled value of rate of requests, diversity index and
probability hit at each of the plurality of routers.
9. The method claimed in claim 5, wherein when the cache router
score of the plurality of edge routers is equal, the priority will
be based on the average end-to-end delay being observed while
serving the requests.
10. The method claimed in claim 1, wherein the source router of the
multi-router network is either a content producer server or at
least one router of the multi-router network.
11. The method claimed in claim 1, wherein each of the plurality of
edge routers sends information comprising value of average
probability hit of the contents and value of end-to-end delay being
observed while serving the requests of the contents.
12. A system for improving network performance by co-operative
caching content in a multi-router information centric network (ICN)
architecture, the system comprising: a plurality of edge routers of
the multi-router network; a central router of the multi-router
network; an ICN manager and/or ICN resolver; atleast one router of
a multi-router network is a source of a plurality of contents; and
an end user, wherein the end user is communicatively coupled with
multi-router network over the ICN.
13. The system claimed in claim 11, wherein the plurality of edge
routers, the central router and the atleast one source router of
multi-router network are comprising the steps of: on-path caching,
wherein a plurality of edge routers of a multi-router network cache
an incoming content from a source router in the return path while
the content is being served to an user of the multi-router network;
and off-path caching, wherein a predefined central router of the
multi-router network caches one or more un-cached contents and one
or more filtered out staled contents of each of the plurality of
edge routers.
14. The system claimed in claim 12, further comprising: the
plurality of edge routers of the multi-router network are assisted
by the pre-defined central router of the multi-router network; and
the incoming content cached at atleast one of the plurality of edge
routers of the multi-router network is based on availability of
cache of the edge router.
15. The system claimed in claim 12, further comprising: means for
computing score of the incoming content to cache at each of the
plurality of edge routers of the multi-router network, wherein the
computed score is used to tag cache lifetime of the content; and
means for replacing stale cached content of the atleast one of the
plurality of edge routers, wherein the replacement is based on the
lower cache lifetime of the staled cache and incoming content.
Description
PRIORITY CLAIM
[0001] This U.S. patent application claims priority under 35 U.S.C.
.sctn.119 to India Application Serial No. 201621001593, filed on
Feb. 15, 2016. The entire contents of the aforementioned
application are incorporated herein by reference.
TECHNICAL FIELD
[0002] The present subject matter described herein, in general,
relates to Information Centric Networks (ICN). Particularly, the
application provides a method and system for cooperative on-path
and off-path caching policy for information centric networks.
BACKGROUND
[0003] Caching in Information Centric Networks (ICN) is an emerging
area of research. Existing prior art provides a detailed
description of ICN including current and future research directions
on various challenges of ICN such as architecture, naming, caching,
security, etc. while certain other existing prior arts speak about
on-path caching and off-path caching on a broad level.
[0004] In an existing prior art, authors have proposed to use
on-path caching with Least Recently Used (LRU) as the cache
replacement policy, while another prior art proposes a Least
Valuable First (LVF) policy employing a utility function that uses
delay, popularity, and age parameters to determine the cached item
to be replaced. It maximizes user gain according to a delay limit,
thereby ensuring delay profile to the users. It is a sub-optimal
policy and does not explain the nature of the requests entering to
the router. A universal on-path caching in which a content metric
based caching and cache replacement policy is used for optimal
network gain has also been discussed and characterization of the
requests and optimal cache size of the ICN router to ensure a
desired cache hit performance has been proposed too.
[0005] It has been observed in existing prior art that strategic
location of the on-path cached routers are even more important than
an efficient cache replacement policy. While most of the existing
research have intended to improve the latency and the overall
network efficiency, using storage efficiency of the cache space for
on-path caching has also been proposed.
[0006] Contrary to on-path caching, off-path caching saves the
contents in a strategic location to improve overall network usage
and avoid cache duplication. Existing art proposes an off-path
caching scheme using the concept of Software Defined Networks
(SDN). It mainly aims to measure the popularity index associated
with each content and based on that it caches at a strategic
location, while ensuring low overhead.
[0007] Prior art also proposes a joint scheme in which Least
Recently Used (LRU) is used for on-path caching in addition to
off-path caching, such that redundant caching of data items are
avoided. However LRU is used in ICN, the scheme is not optimal.
Further, prior art has also proposed using indirect routing or
traffic deflection such that all requests for a particular content
traverse through a particular path. So, in addition to scalability
issues, characterizing the requests at the beginning and then
defining a route does not lead to an optimal solution.
[0008] Caching as an optimization problem has also been
investigated in numerous ways. Existing art has proposed a formal
framework of cache allocation models in ICN and used this framework
to compare the performance of the scheme. It further explains the
need of caching at a router near the user, while using
Mixed-Integer Linear Programming to formulate on-path caching.
Distance of the source as a factor has been discussed while
modelling the optimization problem. Furthermore, an optimal scheme
in which replication of information inside the network is minimized
has been proposed where using four on-line intra-domain cache
management algorithms with different level of automaticity and
comparing them with respect to performance, complexity, execution
time, and message exchange overhead have been suggested.
[0009] From the existing art discussed, it can be clearly observed
that an optimal caching policy is still required for an ICN network
while ensuring minimal duplication of cached content.
[0010] Accordingly, the present system and method describes static
analysis-based efficient elimination of one or more false positives
from static analysis warnings generated during the property
verification.
OBJECTIVES
[0011] In accordance with the present disclosure, an objective of
the disclosure is to provide a method and system for cooperative
on-path and off-path caching policy for information centric
networks (ICN).
[0012] Another objective of the disclosure is to provide an optimal
co-operative caching policy which can be used in campus level or
ISP level ICN caching design and management.
[0013] Yet another objective of the disclosure is to provide a
method for improving the performance of the network in terms of
network usage and delay at both local (on-path edge routers) as
well as at a global (ISP or campus-wide) level.
SUMMARY
[0014] The following presents a simplified summary of some
embodiments of the disclosure in order to provide a basic
understanding of the embodiments. This summary is not an extensive
overview of the embodiments. It is not intended to identify
key/critical elements of the embodiments or to delineate the scope
of the embodiments. Its sole purpose is to present some embodiments
in a simplified form as a prelude to the more detailed description
that is presented below.
[0015] In view of the foregoing, an embodiment herein provides a
system and a method to improve network performance by co-operative
caching content in a multi-router information centric network (ICN)
architecture.
[0016] In one aspect, a computer implemented method to improve
network performance by co-operative caching content in a
multi-router information centric network (ICN) architecture,
wherein the method comprising on-path caching and off-path caching.
In the on-path caching a plurality of edge routers of a
multi-router network cache an incoming content from a source router
in the return path while the content is being served to an user of
the multi-router network and in the off-path caching a predefined
central router of the multi-router network caches one or more
un-cached contents and one or more filtered out staled contents of
each of the plurality of edge routers. Further, the plurality of
edge routers of the multi-router network are assisted by the
pre-defined central router of the multi-router network and the
incoming content cached at atleast one of the plurality of edge
routers of the multi-router network is based on availability of
cache of the edge router. Furthermore, the method comprising
computing score of the incoming content to cache at each of the
plurality of edge routers of the multi-router network, wherein the
computed score is used to tag cache lifetime of the content and
replacing stale cached content of the at least one of the plurality
of edge routers, wherein the replacement is based on the lower
cache lifetime of the staled cache and incoming content.
[0017] In another aspect, a system for improving network
performance by co-operative caching content in a multi-router
information centric network (ICN) architecture. The system
comprising a plurality of edge routers of the multi-router network,
a central router of the multi-router network, an ICN manager and/or
ICN resolver and at least one router of a multi-router network is a
source of a plurality of contents and an end user, wherein the end
user is communicatively coupled with multi-router network over the
ICN. Further, the plurality of edge routers, the central router and
the at least one source router of multi-router network is
configured to perform the steps of on-path caching and off-path
caching. In the on-path caching a plurality of edge routers of a
multi-router network cache an incoming content from a source router
in the return path while the content is being served to an user of
the multi-router network and in the off-path caching a predefined
central router of the multi-router network caches one or more
un-cached contents and one or more filtered out staled contents of
each of the plurality of edge routers. Further, the plurality of
edge routers of the multi-router network are assisted by the
pre-defined central router of the multi-router network and the
incoming content cached at at least one of the plurality of edge
routers of the multi-router network is based on availability of
cache of the edge router. In addition to, the system is configured
to compute score of the incoming content to cache at each of the
plurality of edge routers of the multi-router network, wherein the
computed score is used to tag cache lifetime of the content and
replacing stale cached content of the atleast one of the plurality
of edge routers, wherein the replacement is based on the lower
cache lifetime of the staled cache and incoming content.
[0018] It should be appreciated by those skilled in the art that
any block diagram herein represent conceptual views of illustrative
systems embodying the principles of the present subject matter.
Similarly, it will be appreciated that any flow charts, flow
diagrams, state transition diagrams, pseudo code, and the like
represent various processes which may be substantially represented
in computer readable medium and so executed by a computing device
or processor, whether or not such computing device or processor is
explicitly shown.
BRIEF DESCRIPTION OF THE DRAWINGS
[0019] The embodiments herein will be better understood from the
following detailed description with reference to the drawings, in
which:
[0020] FIG. 1 illustrates a system for improving network
performance by co-operative caching content in a multi-router
information centric network (ICN) architecture according to an
embodiment of the present disclosure.
[0021] FIG. 2 illustrates an Information Centric Networks (ICN)
topology according to an embodiment of the present disclosure.
[0022] FIG. 3 illustrates the functionalities of an Information
Centric Networks (ICN) router according to an embodiment of the
present disclosure.
[0023] FIG. 4 illustrates a network topology with ICN on-path and
off-path caching routers according to an embodiment of the present
disclosure.
[0024] FIG. 5 illustrates a method to improve network performance
by co-operative caching content in a multi-router information
centric network (ICN) architecture according to an embodiment of
the present disclosure
[0025] FIG. 6 illustrates the performance analysis of the proposed
scheme in terms of average delay according to an embodiment of the
present disclosure.
[0026] FIG. 7 illustrates the performance analysis of the proposed
scheme in terms of average cost which is a function of network
usage according to an embodiment of the present disclosure.
[0027] FIG. 8 illustrates the performance of the system in terms of
delay with changing cache miss percentage at different values of
request traffic rate according to an embodiment of the present
disclosure. and
[0028] FIG. 9 illustrates the performance of the system in terms of
cost or network usage with changing cache miss percentage at
different values of request traffic rate according to an embodiment
of the present disclosure.
DETAILED DESCRIPTION OF EMBODIMENTS
[0029] Exemplary embodiments are described with reference to the
accompanying drawings. In the figures, the left-most digit(s) of a
reference number identifies the figure in which the reference
number first appears. Wherever convenient, the same reference
numbers are used throughout the drawings to refer to the same or
like parts. While examples and features of disclosed principles are
described herein, modifications, adaptations, and other
implementations are possible without departing from the spirit and
scope of the disclosed embodiments. It is intended that the
following detailed description be considered as exemplary only,
with the true scope and spirit being indicated by the following
claims.
[0030] The present disclosure relates to a computer implemented
method and system for cooperative on-path and off-path caching
policy for information centric networks (ICN) which can be used for
improving the performance of the network in terms of network usage
and delay at both local (on-path edge routers) as well as at a
global (ISP or campus-wide) level.
[0031] A system 100 for improving network performance by
co-operative caching content in a multi-router information centric
network (ICN) architecture. The system comprising a plurality of
edge routers of the multi-router network 102, a central router of
the multi-router network 104, an ICN manager and/or ICN resolver
106 and at least one router of a multi-router network is a source
of a plurality of contents 108 and an end user 110, wherein the end
user is communicatively coupled with multi-router network over the
ICN.
[0032] ICN envisions a radical shift from existing host-centric
networking of the Internet to a content centric networking. Caching
is an important and integral part of ICN. In ICN, the requested
contents that are downloaded from the original source are stored at
intermediate on-path routers. However, with increase in similarity
in the requests between the edge routers, cache redundancy occurs
at the edge routers leading to decrease in network performance.
[0033] In the preferred embodiment, an overlay ICN architecture in
which the users' requests for contents are resolved by an ICN
Manager is considered. Next, an end-to-end reverse path between the
source and the user is created by the ICN Manager. It is to be
noted that the source can be the original producer server or any
other cache router where the data is currently cached. In the
downloading on-path of the content, each intermediate ICN router
may cache it in its internal cache. Apart from this, the contents
can also be cached in near-by strategically located off-path
caches. Once the content is cached by the routers, ICN Manager is
to be updated so that the ICN Manager keeps track of the sources to
resolve future requests. Based on the updated information, the
future requests for the same content can be directed to the
corresponding ICN routers for delivery.
[0034] In an example, as illustrated in FIG. 2, a network with two
different campuses which are considered to be of smaller versions
of ISP networks. Campuses are divided into various sub-networks
(equivalent to zones or sub-zones under ISP network) in a
department-wide fashion; viz., in Campus 1, there are three
departments, EE: Electrical Engineering, CS: Computer Science and
ME: Mechanical Engineering. Each sub/department-network is having a
local ICN router to serve the requests initiated by the users
associated to it. The department ICN routers are connected with
each other either in a mesh or Local Area Network (LAN) or in a
Wide Area Network (WAN) fashion and are relatively low-end ICN
routers in terms of processing power and caching capacity. Each
Campus is equipped with a Resolver (RE) and a central high-end ICN
router (RC) which can cache the contents in an off-path or off-line
fashion. Multiple campuses can be inter-connected through
border-gateway(s) (G). It is to be noted that operation wise, the
central router `RC` is different from the edge routers as it does
not store any on-path downloaded contents directly and stores
contents that are forwarded from the edge routers for offloading
purpose only. The terms ICN Manager and name resolver has been used
interchangeably hereinafter.
[0035] In the preferred embodiment, the plurality of edge routers,
the predefined central router and the at least one source router of
multi-router network are comprising the steps of on-path caching
and off-path caching. In the on-path caching the plurality of edge
routers cache an incoming content from a source router in the
return path while the content is being served to the user of the
multi-router network. In the off-path caching the predefined
central router of the multi-router network caches one or more
un-cached contents and one or more filtered out staled contents of
each of the plurality of edge routers. Herein the plurality of edge
routers are assisted by the predefined central router and the
incoming content can be cached at atleast one of the plurality of
edge routers based on availability of cache of the edge
routers.
[0036] In another example, as illustrated in FIG. 3, let the finite
cache size at any ICN router i be B-951345943 i. Let
.lamda.-951345943 i be the average rate of requests at router i. On
receiving a request for a content from any of the users, router i
first checks in its local cache. If the content is available, then
it serves the content immediately. In case of a cache miss for the
content, router i forwards the request to the ICN Manager which
performs name resolution. Next, the request to the source of the
content is forwarded by the ICN Manager; serving source can be the
original source of the content or a nearest router which has the
requested content in its cache. While serving the contents passing
through it from the source to the user, any intermediate ICN router
(including router i) can cache the content subject to availability
of cache. If the cache is full, router i needs to replace low
priority older contents with the incoming high priority ones. After
either caching or cache replacement, ICN router updates the ICN
Manager such that future requests can be forwarded to the specific
routers for content delivery. In order to have an intelligent
caching mechanism, router i also needs to characterize the incoming
requests. Let T-951345942 i be the diversity index of the requests
at router i, expressed as:
i = H ( X ) lo g 2 ( .lamda. i ) , ( 1 ) ##EQU00001##
where H(x) is the entropy of the incoming requests X and
log2(A-951345941 i) is the maximum entropy of requests;
0.ltoreq.T-951345941 i .ltoreq.1; T-951345941 i=0: all the requests
are identical and T-951345941 i=1: all the requests are distinct.
To characterize the requests X, the time axis is discretized with a
sample period of T-951345940 t time units. With finite cache size
and large number of contents in the network, while storing, each
content is attached with a cache lifetime tag by the ICN router. If
the content remains un-accessed within its lifetime, then it is
discarded from the cache and the status is updated to the ICN
Manager. This further ensures room for the future requested
contents. A fixed cache lifetime is assumed and each router while
serving requests, need to intelligently decide whether to cache the
requested content or not. In the process, it has to identify the
cached contents which can be replaced to provide room for the new
in-coming contents. For the optimal decision, each router i should
ensure caching of contents which can lead to minimum delay to serve
and maximum probability of cache hit. This is equivalent to solving
the following optimization problem at each router i.
max E [ p i ] , min E [ d i ] , s . t . .DELTA. B i .gtoreq. 0. ( 2
) ##EQU00002##
[0037] This is a multi-objective optimization problem, where
E[p-951345939 i] is the average probability of hit that the ICN
router experiences while serving the requests, E[d-951345939 i] is
the average end-to-end delay being observed while serving the
requests, and .DELTA.B-951345939 i is the available cache size at
any time instant t. The end-to-end delay experienced by the ICN
router is a function of cache search time, name resolution time at
the ICN Manager, and end-to-end packet transmission time which
in-turn is a function of the number of hops the content is
traversed through. Out of these three parameters, cache search time
and end-to-end transmission time can be controlled by the ICN
router; by implementing better cache search algorithms and keeping
the requested content in its own cache or at a near-by router. The
third parameter, i.e., name resolution time at the ICN Manager
depends upon the name size and network size, which are independent
of the caching policy at the ICN router.
[0038] In another embodiment, the system comprising means for
computing score of the incoming content to cache at each of the
plurality of edge routers of the multi-router network and means for
replacing stale cached content of the atleast one of the plurality
of edge routers. The computed score is used to tag cache lifetime
of the content.
[0039] Each ICN router stores the contents which are frequently
asked by the users. This can improve the cache hit probability
E[p-951345937 i]. Further to minimize transmission delay, each
router should cache the contents which are being served from a
distant source, i.e., assign weightage according to the number of
hop distance to the source. To avoid starvation of newly arrived
contents with lower frequency of access, cache lifetime is used.
Therefore, stale cached contents are flushed out, if they are not
accessed for a time T, without impacting the objective of Equation
2.
[0040] Based on the above mechanism, each on-path edge router
computes a content Caching Score CS(k) for each incoming content k
using the following formula and compares this with the scores of
the existing contents in its cache for replacement, in case the
cache is full, i.e., .DELTA.B-951345936 i=0,
CS.sub.i(k)=c.sub.1.times.D.sub.i(k)+c.sub.2.times.F.sub.i(k),
(3)
Where D-951345936 i(k) is the normalized distance in terms of hop
counts between the source and serving router i (normalized with
maximum count possible) which can be obtained with the help of the
Time-to-Live (TTL) value associated with the data packet if
Internet Protocol (IP) stack is used. Something similar to TTL can
also be used in case IP stack is not used by the ICN architecture.
F-951345933 i(k) is the normalized frequency of access of the
content (normalized with the maximum frequency F-951345933 Max of
access of cached contents) and c-951345933 1, c-951345933 2 are
constants. If a cached content is served, then D-951345932 i(k)=0
and
F i ( k ) = F i ( k ) + 1 F Max ##EQU00003##
If a new content (not cached) is served, then
F i ( k ) = 1 F Max and D i ( k ) ##EQU00004##
and D-951345932 i(k) is the normalized hop count for this
content.
[0041] In a multi-router network as illustrated in FIG. 2, multiple
routers solve Eq. (2) independently for intelligent caching. Since
this is a multi-objective optimization problem, the disclosure
proposes to relax one objective out of the two, such that a
practical solutions can be obtained. The system comprises the
central cache router which can be placed strategically in the
network. The new contents fetched from the source but unable to get
cached at the on-path edge routers or the contents which are
getting replaced at the edge routers due to lower CS value can be
cached at the central cache using off-path caching. Hence, Eq. (2)
can be re-written as:
max E [ p i ] , s . t . E [ d i ] .ltoreq. .delta. , .DELTA. B i
.gtoreq. 0 , ( 4 ) ##EQU00005##
Where .delta. is the delay threshold which needs to be maintained
as a part of the Service Level
[0042] Agreement (SLA) by the operator. It is to be noted that Eq.
(4) is a form of Eq. (2), where the second objective is transformed
to a constraint with the help of the new central router. This gives
flexibility in solving the optimization problem.
[0043] Referring FIG. 5, a method 200 to improve network
performance by co-operative caching content in a multi-router
information centric network (ICN) architecture. The method
comprising on-path caching and off-path caching.
[0044] In the preferred embodiment at step 202, in the on-path
caching atleast one of a plurality of edge routers of a
multi-router network caches an incoming content from a source
router in the return path while the content is being served to a
user of the multi-router network. The source router of the
multi-router network is either a content producer server or at
least one router of the multi-router network.
[0045] In the preferred embodiment at step 204, in the off-path
caching a predefined central router of the multi-router network
caches one or more un-cached contents and one or more filtered out
staled contents of each of the plurality of edge routers. Wherein
the plurality of edge routers of the multi-router network are
assisted by the pre-defined central router of the multi-router
network. Each of the plurality of edge routers will send
information consisting the E[p-951345939 I] and E[d-951345939 I]
value of the contents that the plurality of edge routers want to
cache in the central router.
[0046] In the preferred embodiment at step 206, the incoming
content cached at atleast one of the plurality of edge routers of
the multi-router network is based on availability of cache of the
edge router.
[0047] In the preferred embodiment at step 208, the method computes
score of the incoming content to cache at each of the plurality of
edge routers of the multi-router network, wherein the computed
score is used to tag cache lifetime of the content. Further, the
computed content score is based on normalized distance between the
source router and serving router of the multi-router network and
normalized frequency of access of the content. Furthermore, the
cache router score comprising of scaled value of rate of requests,
diversity index and probability hit at each of the plurality of
routers.
[0048] In the preferred embodiment at step 210, the method replaces
stale cached content of the atleast one of the plurality of edge
routers, wherein the replacement is based on the lower cache
lifetime of the staled cache and incoming content. Further, the
caching of the one or more filtered out staled contents from the
plurality of routers and the one or more un-cached contents is
based on state information of the plurality of edge routers, cache
hit probability of the content and average delay being experienced
while submitting a content for caching.
[0049] In the preferred embodiment, when the cache router score of
the plurality of edge routers is equal, the priority will be based
on the average end-to-end delay being observed while serving the
requests.
[0050] In the preferred embodiment, the state information of the
plurality of edge routers include arrival rate and diversity index
of the requests for content. The caching of the one or more
filtered out staled contents from a plurality of routers is based
on a cache router score of each of the plurality of routers.
[0051] In another embodiment of the disclosure, in order to solve
Eq. (4), a heuristic method is proposed which runs at each ICN edge
router independently. In a step of the said method, cache
availability is checked for and if available then the new content
is cached without replacing the existing contents. Else, based on
the content caching score, the router decides whether to replace or
not. If the new content does not get cached then that particular
content is offloaded to the central router for possible off-path
caching. Similarly replaced contents are also offloaded for
off-path caching. Note that in case the new content is fetched from
the central router, then the edge router does not forward it for
off-path caching if it's non-cacheable at the edge router due to
low content caching score.
[0052] In the preferred embodiment, un-cached and replaced contents
of the edge routers are being offloaded to the central router.
Apart from these, the central router does not directly gets any
other content from the source for caching. In practice, the central
router is also of finite cache size. Therefore, the caching
mechanism to be used by the central router should be able to decide
what to cache when multiple contents are arrived for caching from
different edge routers. To achieve this, a framework is proposed in
which the edge routers need to pass on their state information in
terms of arrival rate and diversity index of requests, cache hit
probability being observed and the average delay being experienced
while submitting a content for caching.
[0053] In another example, wherein disclosure provides that when
any router i requests the central router for off-path caching only
if its own cache availability .DELTA.B-951345925 i=0. FIG. 4
illustrates the off-path caching topology in which R-951345925 1, .
. . , R-951345920 n are the edge routers and RC is the central
routers. Therefore, central router RC receives the following
information from all the edge routers.
( .lamda. 1 1 E [ p 1 ] E [ d 1 ] .lamda. 2 1 E [ p 2 ] E [ d 2 ]
.lamda. 3 1 E [ p 3 ] E [ d 3 ] .lamda. i i E [ p i ] E [ d i ]
.lamda. n n E [ p n ] E [ d n ] ) ##EQU00006##
[0054] Each row of this matrix corresponds to the information
received from one edge router. These information can be
piggy-backed suitably with the ICN data packets. Due to the
distributed nature of the network, information received from edge
routers are not fully synchronized. Since each row in the above
matrix corresponds to the local information of a particular router,
getting a global state at the central router is important. Hence,
to have an optimal caching at RC, the following optimization
problem needs to be solved.
max E [ p C ] , s . t . E [ d i ] .ltoreq. .delta. .A-inverted. i ,
.DELTA. B C .gtoreq. 0 , ( 5 ) ##EQU00007##
Where E[p-951345915 C] is the expected probability of hit at RC and
.DELTA.B-951345914 C is the cache availability of the central cache
router at time instant t. Solution of this can be obtained as
follows:
max [ E [ p C ] - a i ( E [ d i ] - .delta. ) ] , .A-inverted. i ,
.DELTA. B C .gtoreq. 0 , ( 6 ) ##EQU00008##
Where a-951345914 i is a Boolean parameter. a-951345913 i=1, if the
content from the i-th edge router is getting cached at the central
router, a-951345913 i=0, otherwise. To select router i, whose
contents can be stored at the central router in the case of
simultaneous requests coming from multiple edge routers, cache
router score si is defined as:
s i = E [ d i ] i , ( 7 ) ##EQU00009##
Where gi is the scaled value of rate of requests, diversity index
and observed probability of hit at router i. In other words, it is
the fraction of non-repetitive requests getting served by any edge
router i:
g.sub.i=E[p.sub.i].times..lamda..sub.i.times..tau..sub.i. (8)
[0055] Central router RC sorts the s-951345911 i values associated
with each edge router i and caches the content of the router i with
the highest s-951345911 i value, i.e., a-951345911 i=1. The central
router keeps on caching the contents of the edge routers with
descending order of s-951345911 i till cache availability. In case,
two routers have same s-951345910 i values, central router gives
priority to the router with higher E[d-951345910 i] value for
caching such that delay at that router can be minimized.
[0056] In another embodiment of the present disclosure, another
method is implemented which is employed at the central router. In
the preferred method, cache hit is checked for; if the new incoming
content from the edge router exists in the cache, then the router
score and observed delay associated with the cached content with
that of the incoming content is updated. In case there is a cache
miss, i.e., if the new content is not available in the central
router, cache availability at the central router is checked for. If
cache is available, then it can be cached with the attached router
score and observed delay. In case there is no cache available, then
whether cache replacement can be performed or not needs to be
checked. In case the router score associated with the new content
is higher than the lowest router score associated with the cached
contents, then new content can replace the cached content with
lowest router score. However, if the router score values are same,
the delay associated is checked for and replaced if the delay
associated with the new content is more than that of the cached
content. In all other cases the incoming content is dropped. To
avoid cache staleness, old contents whose lifetime are expired are
dropped. Once cached, the central router needs to update the ICN
Manager such that future resolution can be directed to the central
router as the potential source. While serving the future requests
coming from the central router, it needs to update the frequency of
access of the requested content by unity. This is to ensure that
any content which is stored at the central router and which is
being requested by the edge routers should be cached for a longer
period. By doing this, delay profile of the edge routers can be
improved.
[0057] In another example, to evaluate the performance of the
proposed scheme, MATLAB-based simulations with limited ICN
functionalities have been conducted. A memory less system in which
individual edge routers receive requests following Poisson process
with rate A for the existing contents stored in the cache as well
as the new incoming contents has been assumed. ICN contents are
picked out from a pool of 100 different types of contents stored in
a source router outside the campus network. Depending upon the
diversity index T selected at a router, ICN contents can be
repeatedly requested by the users. Upon receiving requests, the ICN
routers check their cached contents; in case of a cache-hit, the
routers serve the contents from their caches, else (cache-miss) the
contents are served either from the central cache or the original
source of the contents. The cache-missed contents can either
replace any existing content (cache-miss-replace) or can be
forwarded to the central cache router for caching such that future
requests for the same content can be served without downloading
from the original source. It is to be noted that with Overlay-ICN
(O-ION) type of architectural ICN implementation for each caching
update (cached or replaced or dropped), the ICN routers need to
update the ICN Manager. The ICN Manager is actually the arbitrator
and cache miss/hit corresponding to the ICN routers can only be
measured at the corresponding ICN Manager.
[0058] In the given example, the ICN Manager functionalities with a
global function which is capable of name resolution, source finding
and keeping the cache updates of the routers attached to it (both
the central and edge routers) has been implemented.
[0059] In the given example, in order to obtain statistically
relevant results, sufficient simulation runs with 1000 different
seed values (i.e., different traffic generations) with confidence
interval of 2% of the mean value with a confidence level of 95%
have been performed. For the simulations, a 5-node network with
(both on-path and off-path caching) and without central router
(on-path caching only) have been conducted. In the former case,
cache size of the edge router are assumed to be of 70 units (no
central router) whereas in the latter case the edge routers are
assumed to be of 30 units with a central router of 200 units cache
size. Therefore, total cache size of both the cases are same (350
units each). It has been further assumed that the ICN contents are
of uniform size.
[0060] In the given example, a per-content delay and network
utilization or cost as the two network parameters have been
considered as the two network parameters to measure the system
performance with varying request traffic rate (.lamda.).
[0061] Per-Content Delay signifies how much average delay each ICN
content transmission can encounter. With the increase in request
traffic rate, more number of requests are missed at the edge
routers which in turn get forwarded to the central cache.
Therefore, future requests for the same contents are served from
the central cache instead of downloading from the distant original
source again. Apart from this, cache duplication also gets reduced.
This improves the total network consumption and delay
significantly. This is illustrated in FIG. 6. Network Utilization
or Cost: Considering each ICN data packet transmission in a network
link is associated with certain cost, this parameter explains the
overall network usage or cost by the campus network. This is
illustrated in FIG. 7.
[0062] As illustrated in FIG. 6 and FIG. 7, it has been observed
that with central cache, performance of the campus network is
improved significantly. Reduction in cache duplication by the use
of the central cache router has also been observed.
[0063] In another example, the network performance with different
diversity index of the requests has been observed. However, instead
of plotting the performance over diversity index, plotting has been
done with respect to cache miss percentage. Generated requests are
exactly identical to the cached contents when the cache miss
percentage is zero (i.e., T=0) and generated requests are
completely diverge from the cached contents when the cache miss
percentage is 100 (i.e., T=1).
[0064] Referring FIG. 8 and FIG. 9, the performance of the system
with changing cache miss percentage at different values of request
traffic rate .lamda. has been observed. From FIG. 8, it has been
observed that the delay gets affected with increase in cache miss
percentage or diversity index. In addition to this, with increase
in request rate, average delay also increases. However, as expected
average delay is reduced with central cache router in place as
compared to that without the central cache router. Similarly, FIG.
9 depicts that as the cache miss percentage increases, the cost of
the system also increases. Moreover, with high data traffic flow,
higher number of misses occur which costs the system even more. It
is to be noted that the delay performance (FIG. 8) degrades with
higher request traffic rate with or without a central cache, it is
important to mention that delay performance degradation is much
lower with a central cache even under diverse content requests, and
especially so under heavy request traffic conditions. Similarly,
average costs (FIG. 9) are lower for diverse and heavy traffic
situations with a central cache router. In all it proves that
having a central cache, improves the performance of router clusters
such as in a campus. While these results are for a typical campus,
similar results can be expected in larger Internet by clustering a
group of routers around some criterion such as physical location
and then allocating an off-path cache router at a nearby
location.
[0065] Although implementations of system and method for
integrating data from plurality of sources and to present a
comprehensive view to the consuming applications, it is to be
understood that the specific features and methods are disclosed as
examples of implementations for integrating data from plurality of
sources and to present a comprehensive view to the consuming
applications.
[0066] The illustrated steps are set out to explain the exemplary
embodiments shown, and it should be anticipated that ongoing
technological development will change the manner in which
particular functions are performed. These examples are presented
herein for purposes of illustration, and not limitation. Further,
the boundaries of the functional building blocks have been
arbitrarily defined herein for the convenience of the description.
Alternative boundaries can be defined so long as the specified
functions and relationships thereof are appropriately performed.
Alternatives (including equivalents, extensions, variations,
deviations, etc., of those described herein) will be apparent to
persons skilled in the relevant art(s) based on the teachings
contained herein. Such alternatives fall within the scope and
spirit of the disclosed embodiments. Also, the words "comprising,"
"having," "containing," and "including," and other similar forms
are intended to be equivalent in meaning and be open ended in that
an item or items following any one of these words is not meant to
be an exhaustive listing of such item or items, or meant to be
limited to only the listed item or items. It must also be noted
that as used herein and in the appended claims, the singular forms
"a," "an," and "the" include plural references unless the context
clearly dictates otherwise.
[0067] Furthermore, one or more computer-readable storage media may
be utilized in implementing embodiments consistent with the present
disclosure. A computer-readable storage medium refers to any type
of physical memory on which information or data readable by a
processor may be stored. Thus, a computer-readable storage medium
may store instructions for execution by one or more processors,
including instructions for causing the processor(s) to perform
steps or stages consistent with the embodiments described herein.
The term "computer-readable medium" should be understood to include
tangible items and exclude carrier waves and transient signals,
i.e., be non-transitory. Examples include random access memory
(RAM), read-only memory (ROM), volatile memory, nonvolatile memory,
hard drives, CD ROMs, DVDs, flash drives, disks, and any other
known physical storage media.
[0068] It is intended that the disclosure and examples be
considered as exemplary only, with a true scope and spirit of
disclosed embodiments being indicated by the following claims.
* * * * *