Avian Based Intelligent Algorithm to Provide Zero Tolerance Load Balancer for Cloud Based Computing Platforms

Avian Based Intelligent Algorithm to Provide Zero Tolerance Load Balancer for Cloud Based Computing Platforms

Sivashanmugam G., SP Shantharajah, N.Ch.Sriman Narayana Iyengar
Copyright: © 2019 |Pages: 26
DOI: 10.4018/IJGHPC.2019100104
(Individual Articles)
No Current Special Offers


Artificial intelligence changes the art of solving the computational problems from defined computing structures into disorganized computing structures with the interference of naturally-inspired activities. On this basis, many algorithms were proposed and applied successfully, those results give complete as well as partial solutions for their applications. In this article, the authors consider the same phenomena and the investigation area is a load balancer. The legitimate aim is to bring a zero-tolerance load balancer by applying artificial intelligence techniques. For this, the authors introduced an algorithm named the Eagle Fly algorithm, which is a natural inspired algorithm, completely based on eagle characteristic behavior. From this, the authors examine how tasks are fetched, computed, and server on-demands are supported. This article proves performance metrics received from eagle fly algorithm is good and the results were compared with other existing natural inspired algorithms.
Article Preview


The success behind any computational activity is about finding the right computing method for a given problem and our methodology should be in optimized way. Artificial Intelligence is a factor completely depends upon how a particular problem is solved in an optimized way with naturally inspired ideas. On other hand, nature of load balancer is annoying one and comes under NP hard category. Sensing load balancer turns are impossible because they are time to time varying concepts. Some of them are stated below:

  • Peak load sets and normal load sets on load balancer is unpredictable;

  • Running demand less computational resources are impossible;

  • Executing complexity-based jobs sequentially are a nightmare.

All the time, algorithms are aimed to bring solution as best fit or fit but in load balancer maximum cases, existing algorithms are ended with the worst fit. This is due to complications inside load balancer is standing like a dragon. No algorithm addresses complete solution for load balancer problems because of NP hard, solutions are not up to certainty. Maximum solutions are done with assumptions only. This we called as supportive solution that too obtained by combinational approach. All the services have conceptual approach like peak hour and ideal hour, while talking about web services, no originator can say this is peak hour or ideal hour and maintaining such ratio of computational resources is enough to cross the period. Because, web service is an open platform, any user can enjoy the internet Day by day users involving in online system become multiplicative. Delivering swift response to the user's query is possible if and only if computational resources like application server, searching server, web server are able to meet or having necessary load tolerance capacity, unless systems remain inefficient. Maintaining adequate number of servers is possible only if we know the number of tasks going to execute. This idea is hopeless but trying to solve with various approaches is engineer’s gratitude.


Load Balancer

From the beginning days of distributed system onwards, proposal on load balancing remains suspicion; because the purpose of load balancer and algorithm applied on load balancer is ambivalent (i.e.) purpose and proceeding are opposite to each other which is shown in Figure 1.

Figure 1.

Working model of load balancer


The practice of load balancer is to distribute the load evenly on the available machines but motto of proposed algorithms on load balancer is “keeping resources in idle is worthless”. From the statement conserve above is feasible only if we apply spurious ideas in load balancer. More than 30 years, load balancer struggling hard to complete the assigned job but the matter behind is, whether the solution is optimal or not. The general working design of load balancer is shown in Figure 2.

Figure 2.

Hierarchy of factor affecting load balancer performance


From this design, it is very clear that, the goal of load balancer is not only on distribution of load balancer, it also includes computing techniques and problem-solving techniques to complete activities of request processing. Next, we are going to describe, how these factors affecting the overall performance of load balancer.

After crossing various study materials, we conclude three factors determine performance state of load balancer and they are stated below:

  • Computing techniques followed to establish network connectivity;

  • Kind of algorithm (functionalities, geographical area connectivity) adopted for load balancer;

  • Methodologies to overcome the complications (overloaded, crash among server, on-demand service establishment from private networks) appeared in load balancer.

Complete Article List

Search this Journal:
Volume 15: 2 Issues (2023)
Volume 14: 6 Issues (2022): 1 Released, 5 Forthcoming
Volume 13: 4 Issues (2021)
Volume 12: 4 Issues (2020)
Volume 11: 4 Issues (2019)
Volume 10: 4 Issues (2018)
Volume 9: 4 Issues (2017)
Volume 8: 4 Issues (2016)
Volume 7: 4 Issues (2015)
Volume 6: 4 Issues (2014)
Volume 5: 4 Issues (2013)
Volume 4: 4 Issues (2012)
Volume 3: 4 Issues (2011)
Volume 2: 4 Issues (2010)
Volume 1: 4 Issues (2009)
View Complete Journal Contents Listing