LOAD BALANCING IN CLOUD COMPUTING ENVIRONMENT

What is Load Balancing?




 Load balancing in cloud computing is an important problem and critical component for adequate operations in cloud computing  system and it can also prevent the rapid development of cloud computing.

As many clients over the world are requests for  various services at rapid rate in the recent time. To solve this many load balancing algorithms have built that are  very efficient in allocating the request by selecting various correct virtual machines.

However performance is simulated by using variuos tools Cloudsim,CloudAnalyst ,iFogSim simulator based on various parameters like data processing time and response time etc. and compared with existing  Algorithms.

Many algorithms FCFS, honeybee based load balancing technique, Round Robin, Active Clustering, Active Monitoring Load Balancer, Throttled Load Balancer, Active Clustering, WCAP, JIO, CLBVM, Random sampling have been designed to carry out the client’s request towards the cloud nodes but to ensure effective resource utilization and minimum response time  the term load balancing comes into effect .

Load balancing has two major tasks
 A) Resource allocation or provisioning of resources .
B) Scheduling in distributed environment system.

Lets see the simulation of Algorithm:

  Algorithm: Dynamic Load Management Algorithm

DOI: 10.1109/ICGCIoT.2015.7380567

https://ieeexplore.ieee.org/document/7380567?reload=true&arnumber=7380567 


Input: Number of incoming jobs is   i1, i2 . . . . . . . in
Available VM   j1, j2··· . . . . . . jn
Output: All incoming jobs i1, i2 . . . in are allocated least
Loaded virtual machine among the available j1, j2········ jn.
1.         VmLoadBalancer maintains an index table of VMs and the state of the VM (BUSY/AVAILABLE). At the start all VM’s are available.
2.         DataCenterController then receives a new request.
3.         DataCenterController queries the VmLoadBalancer for the next allocation.
4.         VmLoadBalancer parses the allocation table from top until the first available VM is found or the table is parsed completely.
5.         Check for all current allocation count is less than max length of VM list allocate the VM.
6.         Count the active load on each VM.
7.         Return the id of those VM which is having least load.
8.         The VM Load Balancer will allocate the request to that VM.
9.         If a VM is overloaded then the VM Load Balancer will distribute few of its work to the least loaded VM so that every VM is equally loaded.
10.      If VM not found, VmLoadBalancer returns -1 then DataCenterController queue up the incoming request.
11.      When the VM finishes processing the request and the DataCenterController receives the cloudlet response it notifies the VmLoadBalancer of the VM deallocation.
12.      The Data Center Controller focuses on checking if any waiting requests are there in the queue, it continues from step 3.
13.      Continue from step 2.



Tool-CLOUDANALYST


       The basic components of cloud analyst are:


a)       GUI Packages: It is mainly responsible for the graphical user Interface.

b)   User Base: It models a different group of users that are treated as a single unit in simulation and generates traffic for the simulation.

c)  Internet- It models the Internet traffic routing around the globe by introducing transmission and data transfer delays.

d)   DataCenterController: It generally controls Data center activities.

e) VmLoadBalancer: The Data center controller generally uses a VmLoadBalancer to determine which VM should be assigned the next Cloudlet for processing and models load balancing policies.

f)     Simulation: Simulation accepts requests and then execute request.

g)   CloudAppServiceBroker: This component broker handles the traffic routing between data centers and user base.




Fig.1. User Base Configuration


Fig.2. Application Deployment Configuration


Fig.3. Data Center Configuration


Fig.4. Load Balancing Policy







Fig.5. Output Screen of CloudAnalyst



       Overall Response Time has been improved:

                                       Fig.6. Overall Response Time Summery of Dynamic Algorithm





Fig.7. Overall Response Time Summery of VM-Assign Algorithm









Comments