U.S. patent application number 13/541440 was filed with the patent office on 2012-10-25 for chargeback reduction planning for information technology management.
This patent application is currently assigned to International Business Machines Corporation. Invention is credited to Sandip Agarwala, Ramani R. Routray, Sandeep M. Uttamchandani.
Application Number | 20120271678 13/541440 |
Document ID | / |
Family ID | 43781770 |
Filed Date | 2012-10-25 |
United States Patent
Application |
20120271678 |
Kind Code |
A1 |
Agarwala; Sandip ; et
al. |
October 25, 2012 |
CHARGEBACK REDUCTION PLANNING FOR INFORMATION TECHNOLOGY
MANAGEMENT
Abstract
Minimizing cost chargeback in an information technology (IT)
computing environment including multiple resources. One
implementation involves determining time-based usage patterns and
allocation statistics for a plurality of resources and associated
resource workloads. Using a regression function for determining a
correlation of response time with resource usages and outstanding
input/output instructions for the plurality of resources. Based on
the time-based usage patterns, allocation statistics and the
correlation, deriving an interpolation using positive and negative
integrals to minimize a difference between allocated resource
values and average allocation values. Determining service level
objectives (SLOs) and resource allocation for minimizing cost
chargeback for the resource workloads based on the derived
interpolation.
Inventors: |
Agarwala; Sandip;
(Sunnyvale, CA) ; Routray; Ramani R.; (San Jose,
CA) ; Uttamchandani; Sandeep M.; (San Jose,
CA) |
Assignee: |
International Business Machines
Corporation
Armonk
NY
|
Family ID: |
43781770 |
Appl. No.: |
13/541440 |
Filed: |
July 3, 2012 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
12567582 |
Sep 25, 2009 |
8250582 |
|
|
13541440 |
|
|
|
|
Current U.S.
Class: |
705/7.27 |
Current CPC
Class: |
G06Q 30/0283 20130101;
G06Q 30/04 20130101; G06Q 10/06 20130101 |
Class at
Publication: |
705/7.27 |
International
Class: |
G06Q 10/06 20120101
G06Q010/06 |
Claims
1. A method of minimizing cost chargeback in an information
technology computing environment including multiple resources,
comprising: determining time-based usage patterns and allocation
statistics for a plurality of resources and associated resource
workloads; using a regression function for determining a
correlation of response time with resource usages and outstanding
input/output instructions for the plurality of resources; based on
the time-based usage patterns, allocation statistics and the
correlation, deriving an interpolation using positive and negative
integrals to minimize a difference between allocated resource
values and average allocation values; and determining service level
objectives (SLOs) and resource allocation for minimizing cost
chargeback for the resource workloads based on the derived
interpolation.
2. The method of claim 1, further comprising providing a
recommendation for resource allocation that minimizes chargeback
costs.
3. The method of claim 1, wherein the interpolation comprises an
objective function based on an integral area of the resource
workload, cost rate for each device, and operation zone of each
resource workload.
4. The method of claim 3, wherein the positive integrals comprise
an area of a curve above allocated resource values, and the
negative integrals comprise an area of a curve between average
allocated values and allocated resource values.
5. The method of claim 3, wherein the objective function comprises:
Max j = 0 devices i = 0 workloads A ij + .times. C j .times. SLO i
- j = 0 devices i = 0 workloads A ij - .times. C j .times. SLO i (
2 ) ##EQU00003## wherein A.sub.ij=integral area of resource
workload i at device j, where (A+): area of curve above allocated
resource value, and (A-): area of curve in between average
allocated value and allocated resource value, C.sub.j=Cost rate for
device j, and SLO.sub.i=Operation zone of resource workload i.
6. The method of claim 5, wherein constraints for the objective
function comprise latency of service level objectives for each
resource workload, latency based on outstanding input/output
operations and load of outstanding input/output operations.
7. The method of claim 3, wherein the objective function uses
non-linear optimization that interpolates impact of resource
allocation change on application latency in relation to SLO.
8. The method of claim 1, wherein the regression function
quantifies resource workload latency as a function of number of
outstanding input/output operations.
9. The method of claim 1, wherein the SLOs use randomized bin
packing
10. A system comprising: an evaluation module that evaluate
time-based resource usage patterns and allocation statistics for a
plurality of resources and associated resource workloads; and a
chargeback optimization module that determines a correlation of
response time with resource usages and outstanding input/output
instructions for the plurality of resources, and based on the
time-based resource usage patterns, allocation statistics and the
correlation, derives an interpolation using positive and negative
integrals to minimize a difference between allocated resource
values and average allocation values, and determines service level
objectives (SLOs) and resource allocation for minimizing cost
chargeback for the resource workloads based on the derived
interpolation.
11. The system of claim 10, further comprising an enterprise
network coupled to the system.
12. The system of claim 10, wherein the plurality of resources
comprise: storage devices; latency allocation; at least one server
device; a plurality of switches; and a plurality of applications
executed by the at least one server device.
13. The system of claim 10, wherein a final cost reduction
recommendation is provided by the chargeback optimization module,
wherein the final cost reduction recommendation comprises a time
varying SLO.
14. The system of claim 13, wherein the final cost reduction
recommendation comprises one of a change of a current resource
allocation value and a new resource allocation.
15. The system of claim 13, wherein the final cost reduction
recommendation defines resource allocation that minimizes
chargeback costs.
16. A computer program product for minimizing chargeback costs
comprising a non-transitory computer usable medium including a
computer readable program, wherein the computer readable program
when executed on a computer causes the computer to: determine
time-based usage patterns and allocation statistics for a plurality
of resources and associated resource workloads; use a regression
function for determining a correlation of response time with
resource usages and outstanding input/output instructions for the
plurality of resources; based on the time-based usage patterns,
allocation statistics and the correlation, derive an interpolation
using positive and negative integrals to minimize a difference
between allocated resource values and average allocation values;
and determine service level objectives (SLOs) and resource
allocation for minimizing cost chargeback for the resource
workloads based on the derived interpolation.
17. The computer program product of claim 16, wherein the computer
readable program when executed on the computer further causes the
computer to: provide a recommendation for resource allocation that
minimizes chargeback costs.
18. The computer program product of claim 16, wherein the
interpolation comprises an objective function based on an integral
area of the resource workload, cost rate for each device, and
operation zone of each resource workload.
19. The computer program product of claim 18, wherein the positive
integrals comprise an area of a curve above allocated resource
values, and the negative integrals comprise an area of a curve
between average allocated values and allocated resource values.
20. The computer program product of claim 16, wherein the objective
function uses non-linear optimization that interpolates impact of
resource allocation change on application latency in relation to
SLOs.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This patent application is a continuation patent application
of U.S. patent application Ser. No. 12/567,582, filed on Sep. 25,
2009, the disclosure of which is incorporated herein its entirety
by reference.
BACKGROUND
[0002] 1. Field of the Invention
[0003] The present invention relates generally to an information
technology (IT) environment, and more in particularly, to reducing
chargeback costs in an IT environment.
[0004] 2. Background Information
[0005] With increasing automation of business processes within
enterprises, the demand for information technology (IT)
infrastructure is increasing exponentially, and is now a
significant percentage of the total operating cost of a business.
The capacity, performance, and availability demands of individual
departments within each enterprise are being consolidated from
isolated server-storage silos, to a unified virtualized environment
of servers hosting multiple on-demand virtual machines, with
transparent access to the entire storage subsystem using storage
area networks (SANs). While such consolidation helps management, it
poses challenges for Chief Information Officers (CIOs) responsible
for containing IT costs and regulating usage of the infrastructure
within departments. Chargeback is a process used to regulate IT
costs by charging each department proportionally according to the
resource allocated to it. This fosters efficient use of the
available resources and also makes departments aware of their IT
usage and associated costs.
[0006] A typical enterprise environment includes multiple
departments, each utilizing custom IT applications and IT resource
services. With the advancement of technologies such as
virtualization and multi-core architectures, such IT custom
applications and resource services are deployed in a shared and
consolidated server-storage environment, typically managed by the
enterprise IT department. Resource allocation for the applications
is provided either by humans or resource planners. One example of
resource allocation planner is TotalStorage Productivity Center
(TPC) Storage Area Network (SAN) Planner. The allocation technique
is dependent on the application Service Level Objectives (SLO),
defined in terms of maximum latency, minimum throughput, etc. The
allocation technique may also depend on quality attributes
including no single point of failure, disaster recovery support,
etc. Capacity planning involves utilizing automated tools to
allocate a set of resources for a given set of SLOs. This is
accomplished in two broad steps: first, the resources needed to
achieve SLOs of each customer are determined using workload and
device models (e.g., queuing theory model) and second, the
resources are allocated from available resources using one of the
many multi-dimensional bin packing algorithms. An IT department
keeps track of the usage of these resources and depending on their
usage allocates costs to each department in the form of a
chargeback.
[0007] Depending on the chargeback policies, departments may be
charged, whether or not they use the resources allocated to them.
Although an IT department recovers total operating cost in the form
of chargeback, enterprise as a whole may suffer due to the
opportunity cost associated with the unused resources. System
administrators or IT service providers while performing resource
allocations attempt to achieve one or more of the following goals:
satisfy customers SLOs, optimize the overall utilization of the
resources, accommodate as many customers as possible, maximize
profits and reduce operational costs. Because in general, IT
customers and providers are conservative, and resources are
over-provisioned to handle peak loads. This translates to misuse of
resources and higher chargeback for customers.
BRIEF SUMMARY
[0008] Reducing cost chargeback in an information technology (IT)
computing environment including multiple resources is provided. An
embodiment involves a system including an input module configured
to input network statistics for a plurality of system resources and
a plurality of cost chargeback models. The system further includes
an evaluation module configured to evaluate time-based resource
usage based on the network statistics to result in at least one
resource usage pattern. The system also includes a chargeback
optimization module configured to determine cost reduction
recommendations based on the at least one resource usage pattern
and the plurality of cost chargeback models.
[0009] Another embodiment involves a process wherein resource usage
and allocation statistics are stored for a multitude of resources
and associated cost policies. Then, time-based usage patterns are
determined for the resources from the statistics. A correlation of
response time with resource usages and outstanding input/output
transactions is determined. Based on usage patterns and the
correlation, a multitude of potential cost reduction
recommendations are determined. Further, a multitude of integrals
are obtained based on the potential cost reduction recommendations,
and a statistical integral is obtained based on the statistics. A
difference between the statistical integral and each of the
multiple integrals is obtained and compared with a threshold to
determine potential final cost reduction recommendations. A final
cost reduction recommendation is then selected from the potential
cost reduction recommendations.
[0010] Yet another embodiment involves a computer program product
for reducing cost chargeback in an IT computing environment
including multiple resources. The computer program product
comprises a computer usable medium including a computer readable
program having program instructions. The computer readable program
when executed on a computer causes the computer to perform the
above process.
[0011] Other aspects and advantages of the present invention will
become apparent from the following detailed description, which,
when taken in conjunction with the drawings, illustrate by way of
example the principles of the invention.
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS
[0012] For a fuller understanding of the nature and advantages of
the invention, as well as a preferred mode of use, reference should
be made to the following detailed description read in conjunction
with the accompanying drawings, in which:
[0013] FIG. 1 illustrates a system for reducing cost chargeback in
an IT computing environment including multiple resources, according
to an embodiment of the invention;
[0014] FIG. 2 illustrates an enterprise network including a system
for reducing cost chargeback in an IT computing environment
including multiple resources according to an embodiment of the
invention;
[0015] FIG. 3 illustrates inputs and outputs of a system for
reducing cost chargeback in an IT computing environment including
multiple resources, according to an embodiment of the
invention;
[0016] FIG. 4 illustrates a graph showing positive and negative
integrals used for optimizing chargeback reduction;
[0017] FIG. 5 illustrates a process for reducing cost chargeback in
an IT computing environment including multiple resources, according
to an embodiment of the invention; and
[0018] FIG. 6 illustrates a distributed system, according to an
embodiment of the invention.
DETAILED DESCRIPTION
[0019] The following description is made for the purpose of
illustrating the general principles of the invention and is not
meant to limit the inventive concepts claimed herein. Further,
particular features described herein can be used in combination
with other described features in each of the various possible
combinations and permutations. Unless otherwise specifically
defined herein, all terms are to be given their broadest possible
interpretation including meanings implied from the specification as
well as meanings understood by those skilled in the art and/or as
defined in dictionaries, treatises, etc.
[0020] The description may disclose several preferred embodiments
for reducing cost chargeback in an information technology (IT)
computing environment including multiple resources, as well as
operation and/or component parts thereof. While the following
description will be described in terms of chargeback optimization
for clarity and to place the invention in context, it should be
kept in mind that the teachings herein may have broad application
to all types of systems, devices and applications.
[0021] Reducing cost chargeback in an information technology (IT)
computing environment including multiple resources, is provided.
One embodiment comprises a system for reducing cost chargeback in
an information technology (IT) computing environment, the system
including an input module configured to input network statistics
for a plurality of system resources and a plurality of cost
chargeback models. The system further includes an evaluation module
configured to evaluate time-based resource usage based on the
network statistics to result in at least one resource usage
pattern. The system also includes a chargeback optimization module
configured to determine cost reduction recommendations based on the
at least one resource usage pattern and the plurality of cost
chargeback models.
[0022] The system implements a process wherein resource usage and
allocation statistics are stored for a multitude of resources and
associated cost policies. Then, time-based usage patterns are
determined for the resources from the network statistics. A
correlation of response time with resource usages and outstanding
input/output transactions is determined. Based on usage patterns
and the correlation, a multitude of potential cost reduction
recommendations are determined. Further, a multitude of integrals
are obtained based on the potential cost reduction recommendations,
and a statistical integral is obtained based on the statistics. A
difference between the statistical integral and each of the
multiple integrals is obtained and compared with a threshold to
determine potential final cost reduction recommendations. A final
cost reduction recommendation is then selected from the potential
cost reduction recommendations.
[0023] FIG. 1 illustrates an example block diagram of a system for
reducing cost chargeback in an information technology (IT)
computing environment including multiple resources according to one
embodiment of the invention. The system 100 includes an input
module 110, an evaluation module 120 and a chargeback optimization
module 130.
[0024] FIG. 2 illustrates system 100 as part of an enterprise
system 200 including server devices 1-N 210, switches 1-N 215,
storage units 1-N 220 that include physical memory 221 and logical
memory 222, and client devices 1-N 230. The system 100 functions as
a cost-reduction planner for enterprise systems.
[0025] The system 100 analyzes the historic pattern of resource
usage by different applications and recommends new allocation
strategies that reduce the discrepancies between the actual usage
and allocation. Time-varying SLOs define trends and seasonality of
the resource workload (resource load), allowing adjustments to the
allocation based on the application requirements, for reducing
chargeback. The system 100 provides the ability to make changes in
allocation and SLOs that allow meeting budget requirements of IT
departments, and performing what-if analysis in evaluating cost
savings for different SLOs and provisioning levels.
[0026] FIG. 3 illustrates a process 300 for inputs and outputs of
the system 100 in reducing cost chargeback in an IT computing
environment including multiple resources according to an embodiment
of the invention. The input module 110 inputs resource usage
history data, Service Level Objectives (SLO), allocation data and
resource configuration data. Specifically, in processing block 310,
the input module 110 obtains information from Storage Resource
Management (SRM) tools, including information about average
resource usage, SLO, capacity, allocated resources, and
configuration such as cost policy (for heterogeneous resources).
The evaluation module 320 then analyzes the input information and
provides strategies to reduce application chargeback, including:
changing current allocation values, time-varying SLO and
recommending a new allocation or SLOs.
[0027] In processing block 320 the evaluation module 120 evaluates
time-series chargeback models 340 for performance usage data and
utilizes a regression function to correlate response time with
resource load and number of outstanding IOs. An interpolation is
derived using white-box techniques or by applying known machine
learning algorithms such as CART and M5. The evaluation module 120
uses positive-negative integral functions to optimize resource
allocation that converges to the average application throughput.
These integrals are defined as follows:
[0028] Positive Integral (A+): Area of curve above allocated
value.
[0029] Negative Integral (A-): Area of curve in between average and
allocated value.
[0030] The goal is to minimize the difference between allocated and
average values, such as illustrated by relation (1) below:
Min .intg. 0 t A - .fwdarw. Allocated .apprxeq. Average ( 1 )
##EQU00001##
[0031] The chargeback optimization module 130 achieves optimization
using an objective function such as illustrated by relation (2)
below:
Max j = 0 devices i = 0 workloads A ij + .times. C j .times. SLO i
- j = 0 devices i = 0 workloads A ij - .times. C j .times. SLO i (
2 ) ##EQU00002## [0032] wherein [0033] A.sub.ij=Integral area of
resource workload i at device j, [0034] C.sub.j=Cost rate for
device j, [0035] SLO.sub.i=Operation zone of resource workload i.
[0036] The constraints on the above objective function are: [0037]
latency.sub.i=SLO(resource workload), [0038]
latency.sub.i=f(OutstandingIOS), and [0039] OutstandingIOS=g(Load,
A.sub.ij).
[0040] FIG. 4 shows an example graph 400 illustrating positive and
negative integrals used for optimizing chargeback reduction. The
vertical axis represents the amount of resource used and the
horizontal axis represents the time. The objective function
determines SLOs and allocation that substantially optimizes (i.e.,
minimizes) chargeback for the resource workloads under
consideration. The objective function utilizes non-linear
optimization that interpolates the impact of allocation change on
the application latency in relation to the SLO. The interpolation
uses regression functions that quantify resource workload latency
as a function of number of outstanding IOs.
[0041] Finally in the processing block 330 (FIG. 3), the system 100
provides recommendations including: new allocations, new SLOs, new
schedule and time varying SLOs. The goal of these recommendations
is to minimize the gap between the actual usage and the allocated
resources, thereby reducing chargeback. The `new allocation`
strategy according to an embodiment of the invention recommends a
change in the amount of resource allocation or a change in the type
of resource allocated. For example, moving a workload from a very
high-end server to a low-end server may satisfy its requirements as
well as reduce chargeback. The `new SLOs` strategy according to an
embodiment of the invention recommend a change in the workload SLO.
This is useful in scenarios where SLOs are incorrectly set to
levels that are never attained and results in waste of resource and
higher chargeback.
[0042] The `time varying SLOs` strategy according to an embodiment
of the invention recommends different allocation at different time
of the day or at different month of the year. Typically, SLOs are
defined such that they can handle all the peaks in the workloads.
This results in over-provisioning of resources. Most workloads,
however, consume different amount of resources at different point
of operations. For example, some workloads may consume more
resources during the day. Other workloads may consume more
resources during holiday seasons, etc. The time-varying SLOs
strategy according to an embodiment of the invention adapts the
resource allocation according to the time varying nature of
resource usage. The `new schedule` strategy according to an
embodiment of the invention defers workload processing to non-peak
hours when the demand of IT resources is lower. For example,
`backup` jobs in a datacenter can be executed at night when the
resource utilization is typically low.
[0043] FIG. 5 shows a more detailed process 500 for reducing cost
chargeback by the system 100, according to an embodiment of the
invention. In process block 510, resource usage and allocation
statistics are gathered from resource management tools such as
TotalStrorage Productivity Center (TPC) and Tivoli Application
Dependency Discovery Manager (TADDM). In process block 520,
time-series (time-based) usage patterns are evaluated for the
plurality of resources from the statistics (e.g., calculating
average, peak, median throughput). In process block 530, a
regression function is generated to correlate response time with
resource usages and outstanding input/output instructions for the
plurality of resources. In one implementation, response time is
correlated with resource workload and number of outstanding
input/output operations (IOs). In process block 540, based on usage
patterns and the correlation, a plurality of potential cost
reduction recommendations is obtained. In one implementation,
multi-strategy optimization is performed using randomized
bin-packing.
[0044] Sub-processes of block 540 involve blocks 550 and 560,
wherein in block 550, a plurality of integrals based on the
plurality of potential cost reduction recommendations is
determined, and a statistical integral based on the statistics is
determined. In one implementation, positive and negative integrals
are calculated in relation to averaged and allocated
throughput.
[0045] The positive integrals may have values above a current
resource allocation, and the negative integrals may have a value
between an average resource allocation value and the current
resource allocation value.
[0046] In block 560, the allocations are modified using successive
constraint relaxation. Then, a difference between the statistical
integral and each of the plurality of integrals is compared with a
threshold to determine potential final cost reduction
recommendations. Specifically, in process block 570 it is
determined if the variance between the integrals is greater than a
threshold. If not, then in block 580 results are returned. These
results include one or more strategies (discussed in the previous
paragraph.) Otherwise, in block 590 the historical time series
usage window is divided into smaller intervals and recursively the
process is repeated. Accordingly, a final cost reduction strategy
(recommendation) is selected from the potential cost reduction
strategies.
[0047] The final cost recommendation may comprise a time varying
service level objective. The final cost recommendation may comprise
one of a change of a current resource allocation value and a new
resource allocation. The final cost reduction recommendation may
comprise a new schedule for workload execution, or may define
resource allocation that minimizes chargeback costs.
[0048] FIG. 6 illustrates a distributed system 600 according to one
embodiment of the invention, comprising a distributed network
including a plurality of distributed enterprise centers 610 1-N and
the chargeback optimization system 100. In this embodiment, the
distributed enterprise centers 610 each use the chargeback
optimization system 100 for reducing cost chargeback as described
herein.
[0049] The embodiments of the invention can take the form of an
entirely hardware embodiment, an entirely software embodiment or an
embodiment containing both hardware and software elements. In a
preferred embodiment, the invention is implemented in software,
which includes but is not limited to firmware, resident software,
microcode, etc.
[0050] Furthermore, the embodiments of the invention can take the
form of a computer program product accessible from a
computer-usable or computer-readable medium providing program code
for use by or in connection with a computer, processing device, or
any instruction execution system. For the purposes of this
description, a computer-usable or computer readable medium can be
any apparatus that can contain, store, communicate, or transport
the program for use by or in connection with the instruction
execution system, apparatus, or device.
[0051] The medium can be electronic, magnetic, optical, or a
semiconductor system (or apparatus or device). Examples of a
computer-readable medium include, but are not limited to, a
semiconductor or solid state memory, magnetic tape, a removable
computer diskette, a RAM, a read-only memory (ROM), a rigid
magnetic disk, an optical disk, etc. Current examples of optical
disks include compact disk-read only memory (CD-ROM), compact
disk-read/write (CD-R/W) and DVD.
[0052] I/O devices (including but not limited to keyboards,
displays, pointing devices, etc.) can be connected to the system
either directly or through intervening controllers. Network
adapters may also be connected to the system to enable the data
processing system to become connected to other data processing
systems or remote printers or storage devices through intervening
private or public networks. Modems, cable modem and Ethernet cards
are just a few of the currently available types of network
adapters.
[0053] In the description above, numerous specific details are set
forth. However, it is understood that embodiments of the invention
may be practiced without these specific details. For example,
well-known equivalent components and elements may be substituted in
place of those described herein, and similarly, well-known
equivalent techniques may be substituted in place of the particular
techniques disclosed. In other instances, well-known structures and
techniques have not been shown in detail to avoid obscuring the
understanding of this description.
[0054] Reference in the specification to "an embodiment," "one
embodiment," "some embodiments," or "other embodiments" means that
a particular feature, structure, or characteristic described in
connection with the embodiments is included in at least some
embodiments, but not necessarily all embodiments. The various
appearances of "an embodiment," "one embodiment," or "some
embodiments" are not necessarily all referring to the same
embodiments. If the specification states a component, feature,
structure, or characteristic "may", "might", or "could" be
included, that particular component, feature, structure, or
characteristic is not required to be included. If the specification
or claim refers to "a" or "an" element, that does not mean there is
only one of the element. If the specification or claims refer to
"an additional" element, that does not preclude there being more
than one of the additional element.
[0055] While certain exemplary embodiments have been described and
shown in the accompanying drawings, it is to be understood that
such embodiments are merely illustrative of and not restrictive on
the broad invention, and that this invention not be limited to the
specific constructions and arrangements shown and described, since
various other modifications may occur to those ordinarily skilled
in the art.
* * * * *