A System to Maximize Profit with QOS in Multiservice

ISSN 2348–2370
Vol.06,Issue.04,
June-2014,
Pages:209-213
www.semargroup.org
A System to Maximize Profit with QOS in Multiservice Networks
RENAMALA LOHITH KUMAR1, N. SANTHOSH RAMCHANDER2, N. SATHISH KUMAR3
1
PG Scholar, Dept of CSE, S.V.S Group of Institutions, Warangal, A.P, India, Email: [email protected].
2
Asst Prof, Dept of CSE, S.V.S Group of Institutions, Warangal, A.P, India, Email: [email protected].
3
Professor, Dept of CSE, S.V.S Group of Institutions, Warangal, A.P, India, Email: [email protected].
Abstract: In this paper, we introduce a pricing model that ensures efficient resource allocation that provides guaranteed quality
of service while maximizing profit in multiservice networks. Specifically, a dynamic allocation policy is examined that relies
on online measurements while each service class operates under a probabilistic bound delay constraint. We formulate our
generalized optimization algorithm based on the notion of a “profit center” with an arbitrary number of service classes, linear
revenue and nonlinear cost functions and general performance constraints. To ensure the resource constraint is satisfied, we
incorporate adaptive resource bounds to guide the search. We present a rigorous analysis of the properties of the policy that
provides insights into its workings as well as its sensitivity to various parameters. Finally, its performance is evaluated through
an extensive numerical study.
Keywords: Nonlinear Resource Allocation Problem; Adaptive Resource Bounds; Particle Swarm Optimization; Mathematical
Programming.
I. INTRODUCTION
Recent technology advances have led to dramatic
changes in the communications arena. The use of fiber
optics and the increased performance of integrated circuits
have brought to the forefront diverse types of networks,
such as broadband, wireless ad hoc and mesh networks,
and next generation cellular systems. However, these new
technologies are not sufficient by themselves to guarantee
business success. Added value and service differentiation
need also be considered, in order for the service providers
to be profitable. Hence, the trend towards networks
providing some degree of value added services has
emerged. Specifically, Service Oriented Networks is an
evolving architecture that would allow for a priced based
differentiated choice of network services. Along the same
lines is the triple play network architecture, a user-centric
approach in which customers are confronted with a variety
of applications like Voice over IP, IPTV and Video on
Demand and high speed internet services.
The next generation Internet will provide advanced
services, such as Quality of Service (QoS) guarantees, to
users and their applications. As a result of, these
enhancements, it is expected that service providers will
face an increasing number of users as well as a wide
variety of applications. Under these demanding conditions,
network service providers must carefully provision and
allocate network resources (e.g. bandwidth, buffer size,
CPU capacity) for their customers. Provisioning is the
acquisition of large end-to-end network services
(connections) over a long time scale. In contrast, allocation
is the distribution of these provisioned services (via
pricing) to individual users over a smaller time scale.
Determining the optimal amounts to provision and allocate
remains a difficult problem under realistic conditions.
Service providers must balance user needs in the short term
while provisioning connections for the long term.
Furthermore, this must be done in a scalable fashion to
meet the growing demand for network services, while also
being adaptable to future network technologies.
Fig.1. Depiction of the proposed framework: traffic is
divided into two categories; deterministic constraint
and flexible constraint services. The system allocates
the excess resources to the latter set.
This paper presents a modified particle swarm
optimization (PSO) algorithm for engineering optimization
problems with constraints. PSO is started with a group of
feasible solutions and a feasibility function is used to check
if the newly explored solutions satisfy all the constraints.
Copyright @ 2014 SEMAR GROUPS TECHNICAL SOCIETY. All rights reserved.
RENAMALA LOHITH KUMAR, N. SANTHOSH RAMCHANDER, N. SATHISH KUMAR
All the particles keep only those feasible solutions in their
configuration to satisfy the QoS requirements of multiple
memory. Several engineering design optimization
workflows simultaneously.
problems were tested and the results show that PSO is an
efficient and general approach to solve most nonlinear
Since multiple services are often hosted by the same
optimization problems with inequity constraints. In this
server in ASBS, and the services in the same server
paper, we propose a service pricing model that ensures
compete for the limited available resources of the server,
efficient allocation of resources in a dynamic manner in the
such as CPU-time, memory and network bandwidth,
aforementioned multiservice networks. The scheme
different resource allocations will result in different QoS in
requires close on-line monitoring of the incoming traffic.
runtime. In addition, the service compositions occur
We assume a Fractional Brownian Motion traffic model,
dynamically in runtime and with the resource status of
because of its ability to adequately capture characteristics
servers dynamically changing. Thus, allocating the
of real network traces, such as self-similarity and the
resources of each server to its services for successfully
presence of heavy tailed marginal distributions.
satisfying the QoS requirements of dynamic multiple
workflows and resource status is needed to satisfy the
Optimal resource allocation is also studied in [2], [10].
overall QoS requirements of the workflows in ASBS. In
Specifically, Peng et al. propose a measurement-based
order to develop SBS with the capability of adaptive
resource allocation scheme based on a linear pricing model
resource allocation to satisfy the QoS requirements of
and average queue delay guarantees. This scheme has the
multiple workflows, the following challenges need to be
disadvantage of not being scalable to large number of
addressed:
service classes. Moreover, average queue delay is not
always an appropriate QoS constraint. In [10], they
C1. Situation Awareness: An SBS must be aware of
perform maximization over a utility function provided
changing situations in dynamic runtime environments. An
from the network users and resources are shared based on
efficient and reliable monitoring scheme is needed for
the solution of that optimization problem. In, the authors
collecting the relevant contextual data of situation, such as
study the problem of resource allocation with dynamic
dynamic workflow compositions, the number of servicepricing in which the network administrator controls the
requests for each workflow, QoS requirements, priorities
price of the resources that users demand; based on the
of workflows, and available system resource as well as
demand the prices are dynamically changed over different
various relevant environmental attributes.
time periods so as to maximize the revenue of the
administrator. Finally, measurement-based resource
C2. Context analysis and QoS estimation: An SBS must
allocation has also been studied in different contexts in
be able to efficiently analyze the relationship between
[11]. The remainder of this paper is structured as follows.
resource allocation and QoS of workflows. An automated
Challenges for Adaptive Resource Allocation are described
analysis of the monitored context data should generate
in Section II, Current State of art of Resource Allocation in
good estimates of the expected QoS of workflows.
Section III. Some numerical results illustrating the model’s
performance are presented in Section IV, while some
C3. Optimal resource allocation: An SBS must be able to
concluding remarks are drawn in Section V.
efficiently find an optimal resource allocation in runtime to
satisfy the QoS requirements of multiple workflows with
dynamic situations. If some of the QoS requirements of
II.CHALLENGES FOR ADAPTIVE RESOURCE
workflows cannot be satisfied, then the system must be
ALLOCATION
A major challenge for the Internet ware systems to
able to adaptively sacrifice some requirements according
satisfy various application requirements is to manage the
the contextual data mentioned in C1)
quality of service (QoS) in runtime due to the dynamic,
loosely coupled, and compositional nature of SOA. Many
C4. Implementation of resource adaptation: Once an
studies have identified important QoS features of SOA
optimal resource allocation is determined, SBS must be
systems, such as throughput, timeliness and security, which
able to adaptively change the resource allocation to its
are directly affected by the limitation of system resources.
services in runtime due to dynamic situation.
It is noted that the system resource considered here is for
large-scale aggregation of distributed computing resources
C5.Efficiency and scalability: The processes of collecting
working together over the Internet as a tremendous virtual
and analyzing contextual data to find an optimal resource
computer. For example, computing grids and clouds are
allocation and adapt resource allocation must be efficient
vast resource pools for on-demand requests over the
and scalable
Internet. In order to manage the QoS for such systems, an
effective resource allocation technique for dynamically
III. CURRENT STATE OF ART OF RESOURCE
scalable distributed system resources is needed. In order to
ALLOCATION
achieve this goal, the SBS needs the capabilities of
Recent studies have provided numerous approaches and
monitoring the changing system status, analyzing and
models to facilitate resource management efficiently
controlling system QoS features, and adapting its service
supporting not just global mobility of mobile users but also
International Journal of Advanced Technology and Innovative Research
Volume. 06,IssueNo.04, June-2014, Pages:209-213
A System to Maximize Profit with QOS in Multiservice Networks
enabling agile service creation and access agnosticism.
workload, 4) allocating system resources to VMs in cloud,
These studies are been done keeping in account the
5) pricing the usage of resources and prioritizing the
perspective of both client and network. Some studies have
resource allocation, and 6) keeping track of the execution
also emphasized on the usage of Multiple Stream
progress of the service requests and maintaining the actual
Reservation Protocol (MSRP) over Resource Reservation
usage of resources. This approach supports negotiation of
Protocol (RSVP) in order to provide real-time services by
QoS between users and providers to establish service-level
securing the bandwidth necessary to cater seamless service
agreements (SLA) and allocation of system resources to
to mobile users. We describe next a framework that
meet the SLAs. Multi-layered resource management
extends the more traditional network utility maximization
(MLRM) architecture using standard-based middleware
to include presence and distance-awareness in virtual
technology was developed for enterprise distributed realcollaboration environments (VCEs) and similar social
time embedded (DRE) systems.
settings and goes beyond our previous work on SONs. In
principle, multiple computing and communication
This architecture supports dynamic resource management
resources might have to be allocated, including CPU cycles
to optimize and reconfigure system resources at runtime in
and memory to the virtual machines (VM) running the
response to changing mission needs and resource status.
required applications (or, more generally, the virtual
With the dynamic resource allocation, a DRE system can
machine that hosts the virtual environment that users
provide QoS for critical operations under the overloaded
belong to). Our proposed optimization framework consists
and resource constrained environments. In, a constraintof the following two components: (i) A variation of the 2D
programming-based approach is presented to solve
Knapsack optimization problem that decides which VMs
resource allocation problem in real-time systems. The
would be placed for execution according to the available
problem of assigning a set of preemptive real-time tasks in
resources and their pricing parameters, (ii) a nonlinear
a distributed system is formulated as a constraint
programming problem that allocates the excess resources
satisfaction problem (CSP) with allocation, resource and
to the VMs placed for execution.
timing constraints. First, the CSP is solved using constraint
programming techniques to satisfy the allocation and
Development of QoS and resource management
resource constraints. Then, the solution is validated for
middleware has been studied in various research areas,
timing constraints through Logic-based Benders
such as real-time systems, distributed object-oriented
decomposition. In a decentralized local greedy mechanism
systems, web services, grid computing and cloud
for dynamic resource allocation in web service applications
computing. A workflow-based computational resource
was presented. In this mechanism, software agents are
broker is presented for grid computing environment. The
generated to buy and sell network services and resources to
main function of the resource broker is to monitor the
and from each other.
available resources and match workflow requirements to
available resources over multiple administrative domains.
In, a market based resource allocation for web services
The resource broker provides a uniform interface for
in a commercial environment was presented. In this
accessing available system resources of computing grid via
approach, service providers employ a cluster of servers to
users’ credentials. An open-source toolkit has emerged as a
host web services, and service consumers pay for service
standard middleware for resource management in grid
usage with QoS requirements, such as waiting time or
computing environment provides Web Service Grid
response time. A framework was developed for evaluating
Resource Allocation and Management (WS GRAM)
the effect of particular resource allocation in terms of
Protocol in which a set of web services are designed to
performance and average revenue earned per unit time. A
support APIs for requesting and using grid system
heuristic algorithm was also developed for making
resources also provides Monitoring and Discovery Service
resource allocation decisions in order to maximize revenue.
(MDS) that supports a common interface for collecting
The QoS of networks has been investigated for providing
contextual information on system resources, such as
prioritized services by efficient resource allocation
available processors, CPU load, network bandwidth, file
techniques through labeling, scheduling and routing
system information, storage devices, and memory in
mechanisms. In, an optimal resource allocation and pricing
computing grids.
scheme for next generation multiclass networks was
presented. In this scheme, an optimization problem is
A market-based autonomic resource management
formulated based on a nonlinear pricing model, whose
approach in cloud computing environment was developed,
solution ensures satisfaction of network delay constraints
in which Service-Level Agreement Resource Allocator
and efficient resource allocation in dynamic multiservice
provides the interfaces between the cloud service provider
networks. In, an energy-efficient radio resource allocation
and external users/brokers for 1) monitoring users’ service
approach based on game theory for wireless networks was
requests and QoS requirements, 2) monitoring the
presented.
availability of system resources, 3) examining service
requests to determine whether to accept the requests
The study shows that the game-theoretic approach for
according to resource availability and processing
resource allocation is useful for energy-constrained
International Journal of Advanced Technology and Innovative Research
Volume. 06, IssueNo.04, June-2014, Pages:209-213
RENAMALA LOHITH KUMAR, N. SANTHOSH RAMCHANDER, N. SATHISH KUMAR
wireless networks. In, a resource allocation scheme for the
Hence, we need a new approach to dynamic resource
wireless multimedia applications was presented. In this
allocation for SBS and address the challenges.
scheme, an optimization problem is formulated with
fairness constraints for maximizing system capacity and
IV. PERFORMANCE EVALUATION
resource utilization. The solution of this problem yields the
In this section, we evaluate our pricing model in the over
optimal allocation of sub-channel, path and power. A
provisioned case with a numerical case study. It is assumed
heuristic algorithm to solve the optimization problem was
that there are two types of service classes and the profit
also presented to ensure that the adaptive resource
function becomes:
allocation is performed efficiently in runtime. In, a
resource allocation approach for embedded multimedia
systems using heterogeneous multiprocessors were
(1)
presented, in which optimal resources are allocated to each
application to meet its throughput requirement.
Where
Synchronous Dataflow Graphs (SDFG) is used to model
multimedia applications with time and resource constraints
(2)
and to find the optimal resource allocation.
QoS estimation according to the system activities and
resource status has been studied in many ways. A QoS
model of a router with feedback control that monitors the
state of resource usage and adaptively adjusts parameters
of traffic admission control to estimate QoS and resource
utilization was presented. Batch Scheduled Admission
Control (BSAC) method to predict service delay for high
priority jobs in Internet-type networks was presented. A
regression-based model for dynamically provisioning
resource demand to deliver given QoS expectation was
presented. An adaptive model for the tradeoff between
service performance and security in service-based
environments was presented. This model can be used to
adjust security configurations to provide sufficient
protection and satisfy service performance requirements
with limited system resources. For adaptive resource
allocation for QoS management in SBS, application level
differentiated services were introduced to control QoS for
different classes of service consumers. When the system
resources are limited, fewer resources are allocated for
normal consumers, and most resources are reserved for
satisfying premium consumers’ expected QoS. Feedback
controlled web services were developed to adjust QoS to
meet the most important performance when resources are
limited and consumers’ required performance cannot be
fully satisfied.
In, an integrated QoS management approach in SBS in
order to satisfy user’s QoS requirements by providing
differentiated system resources and priority of workflows
was presented. In, a general methodology for developing
SBS with QoS monitoring and adaptation was introduced.
This methodology supports monitoring of the changing
system status, analysis and control of tradeoffs among
multiple QoS features, and adapting its service
configuration to satisfy multiple QoS requirements
simultaneously. Existing resource allocation approaches
cannot
support
dynamically
changing
runtime
environments in SBS, such as workflow composition, QoS
requirements, workflow priorities and resource status.
Hence, we have to solve the optimization problem:
(3)
The parameters of the profit function used in the study
are shown in Table I. Its concavity over both arguments is
shown graphically in Fig.1, while over the first argument
in Fig.2, by substituting φ2 = 1− φ1. In Tables II, III the
optimal solution is shown when the arrival rate and the
price coefficients are varied. In TableII it can be seen that
with equal arrival rates and all the other parameters the
same, the optimal solution allocates the resource equally
amongst the two classes, as expected. On the other hand,
the class with the higher arrival rate is allocated a larger
portion of the resources, especially if the system is not too
stressed (see rows 2 and 3 in the Table). In that situation
the profit does not also fluctuate much. Finally, when the
system becomes stressed (last row in the Table) the class
with higher arrival rate gets a higher proportion, but the
overall profit for the provider decreases substantially, since
violations of the SLA occur more often and therefore a
large cost is incurred.
TABLE I: Parameters for Two Different Classes
In Table III, the price coefficient varies, while all other
parameters are held fixed (see Table I). Again, with equal
prices we obtain equal allocations, while the allocation of
resources exhibits a strong sensitivity to the price ratio
p1/p2.
International Journal of Advanced Technology and Innovative Research
Volume. 06,IssueNo.04, June-2014, Pages:209-213
A System to Maximize Profit with QOS in Multiservice Networks
V. CONCLUSION
TABLE II: Changing the Arrival Rates
In this paper, we have studied a pricing scheme for next
generation multiservice networks. An optimization
problem based on a nonlinear pricing model was
formulated, whose solution yields the optimal resource
allocation in a network/service node, given the QoS
requirements of each service class that the network element
serves. Our non-linear pricing model responds well to
changes of the characteristics in the input traffic, pricing
parameters and QoS requirements. Further, the resulting
convex optimization problem can easily and efficiently be
solved using standard iterative methods and hence the
proposed modeling framework approach is scalable to any
number of service classes.
Fig.1. Concavity of our Utility Function as a function of
φ1, φ2.
Fig.2. Concavity of our Utility Function as a function of
φ1.
TABLE III: Changing the Pricing Factor pi
VI. REFERENCES
[1] Michael G. Kallitsis, George Michailidis, Michael
Devetsikiotis, “Pricing and Optimal Resource Allocation in
Next Generation Network Services”, pp 389-403, 2014.
[2] Xu P., Devetsikiotis M., Michailidis G., Profit-oriented
Resource Allocation Using Online Scheduling in Flexible
Heterogeneous Networks. Telecommunication Systems, pp
289-303, 2006.
[3] Xu P., QoS Provisioning and Pricing in Multiservice
Networks: Optimal and Adaptive Control over
Measurement-based Scheduling. PhD Dissertation, North
Carolina State University, 2005.
[4] Oltsik J., Web Services Meet the Network. [Online.]
Available: http://library.theserverside.com/detail/RES/11
42273810 440.html.
[5] Sun Q., Pleich R., Sauerwein R., On-Line Measurement
and Analysis of Fractional Brownian Traffic. Proceedings
of the IEEE Conference on High Performance Switching
and Routing, pp 395-400, 2000.
[6] Fonseca N., Mayor G. S., Neto C. A. V., On the
Equivalent Bandwidth of Self-Similar Sources. ACM
Transactions on Modeling and Computer Simulation, Vol.
10, No. 2, pp 104-124, April, 2000.
[7] Norros I., The Management of Large Flows of
Connectionless Traffic on the Basis of Self-similar
Modeling. In Proceedings of IEEE International
Conference on Communications, pp 451-455.
[8] Courcoubetis C., Weber R., Pricing Communication
Networks. John Wiley & Sons, 2003.
[9] Rardin R. L., Optimization in Operations Research.
Prentice Hall, 1998.
[10] Kalyanasundaram S., Chong E. K., Shroff N. B.
Optimal Resource Allocation in Multiclass Networks with
User Specified Utility Functions. The International Journal
of Computer and Telecommunications Networking, pp
613-630, April, 2002.
[11] Chandra A., Gong W., Shenoy P. Dynamic Resource
Allocation for Shared Data Centers Using Online
Measurements. In Proceedings of ACM/IEEE Intl
Workshop on Quality of Service, pp 381-400, 2003.
[12] Knightly E., Shroff N., Admission Control for
Statistical QoS: Theory and Practice. IEEE Network 13(2),
pp 20-29, 1999.
International Journal of Advanced Technology and Innovative Research
Volume. 06, IssueNo.04, June-2014, Pages:209-213