Aims and Objective
The cloud computing is a distributed internet based paradigm,
designed for remote sharing and usage of different resources and
services with high reliability over the large networks
Load balancing in cloud is to balancing load among resource
to obtain resource utilization, maximum throughput;
minimum response time and overhead should be avoided
Dynamic load balancing algorithms distribute the work
among processors during the execution of the algorithm
Literature review of different mechanisms and algorithms
proposed for load balancing in cloud computing.
To study the advantages and flaws of various load
balancing algorithms to identify the problem in load
balancing in cloud computing.
To propose more efficient algorithm for load
balancing to maximize performance, reliability,
scalability and stability in cloud computing.
• monitor resource utility over resource pool
• distribute available resources among severalVMs
• chance of performance degradation due to a large number of
resources employed in frequent dynamic migration
• based on cloud portioning.
• categories idle, normal and overloaded on the basis of load degree
• method of selecting range for load degree has been left
Game theory based
• the least loaded virtual machine for load transfer are selected
• the high migration cost is optimized.
• chance of inefficient service scheduling due to large no. ofVMs and
frequent service requests in the data centre
A genetic algorithm
• Using principle of Ant Colony Optimization.
• disperse a group of tasks evenly on idle nodes using artificial ants.
• convergence speed can be further improved in this system.
An inverse artificial
• finds theCPU utilization, required and available memory for eachVM.
• compares the available resources with required resources, if required resources
are available then proceed further otherwise discard the request
• this mechanism lacks in scalability.
Two phase based load
• more efficient as compared to other algorithms.
• Load agent, channel agent and migration agent.
• can be improved by reducing communication overhead between migration
agent and channel agent.
Based Load Balancing
• It may cause delays, compromised efficiency and less portability.
• There must be some comparison method to allocate resources on
no specific mechanism
to deal with many job
requests at a time
• Self destroy messages might cause extra communication increasing
• Simplicity, reliability and efficiency of the algorithm are affected if
communication overhead is not resolved.
channel agent for self-
• Maintenance of tables causes memory space overhead and affects
the performance of the algorithm by reducing the available
Channel Agent has to
for load balancing
Efficient Decentralized Load Balancing
Algorithm in cloud computing
based on the
User is the task request from the
clients to the cloud
Sequencer will sequence the task
requests from client so that task
waits in the queue for minimum
Load agent is responsible to
transfer the user request to theVM
in the cloud pool
Load Balancer will calculate the
used memory, CPU utilization and
response time of eachVM and
compare it with threshold value.
This work contributes in
two ways; first by providing
a sequencer ,incoming user
requests can be entertained
in more appropriate way.
second load balancer
calculate load status of all
VMs to transfer requested
task to normalVM more
Desired results can be
by implementing this
There is need to implement this work to get desired results
and to resolve more problems regarding load balancing.
Virtualization is the key concept of cloud computing, ifVMs are
located far from one another, there must be some mechanism
to minimize their service time.
More improved algorithms can be designed to provide more
reliability and scalability in load balancing in cloud computing.
A. Singh, D. Juneja and M. Malhotra (2015) ‘Autonomous Agent Based Load Balancing Algorithm in Cloud
Computing’, in proc. International Conference on Advanced ComputingTechnologies and Applications (ICACTA)
Procedia Computer Science, 45,pp. 832-841.
Liu, X. Jin andY.Wang (2005) ‘Agent-Based Load Balancing on homogeneous Minigrids: Macroscopic Modeling
and Characterization’, IEEETransactions on Parallel and Distributed Systems,Volume 1 6, NO.6.
M. Randles, D. Lamb, and A.Taleb-Bendia (2010) ‘A comparative study into distributed load balancing algorithms
for cloud computing’, in Proc. IEEE 24th International Conference onAdvanced Information Networking and
Applications, Perth, Australia. pp. 551-556.
S.C.Wang, K.Q.Yan, W.P.Liao and S.S.Wang (2010) ‘Towards a Load Balancing in a three-Level Cloud Computing
Network’, In Proc. ICCSIT, pp.108-113.
S. Osman, D. Subhraveti, G. Su and J. Nieh (2002) ‘The design and implementation of ZAP: a system for
migrating computing environments’, ACM SIGOPS Oper. Syst. Rev. 36(SI), 361–376.
Y.Xu, L.Wu, L. Guo, Z.Chen, L.Yang and Z.Shi (2011) ‘An Intelligent Load Balancing AlgorithmsTowards Efficient
Cloud Computing’, In Proc. AAAIWorkshop, pp. 27-32.