Context Switching In Clouds: Analysis And Enhancements

2013

visibility

…

description

9 pages

link

1 file

Abstract

This paper presents an methodology to improve the context switch overhead in cloud computing. As migration of jobs in the cloud makes context switching overhead a challenging task which increases the communication costs of cloud and results in the low utilisation. To succeed this objective, dynamic context switching parallel algorithm is purposed in this paper. In order to enhance the utilisation and computing speeds of cloud an improved strategy is used which uses dynamic context switching parallel algorithm. Purposed methodology makes changes in the decision making process of the context switch parallel algorithm dynamically considerably reducing the context switch overheads. For execution of purposed approach, a cloud simulation environment is established which significantly reduces the waiting time of jobs during execution.

Figures (7)

Fig. 1 is showing the life cycle of parallel jobs in distributed environment. Firstly clients submits their jobs for execution, ifno subcloud is free then jobs are queued after this phase node filtering will be done that will detect the currently active nodes by checking the status of all nodes. After node filtering load balancing algorithm will come into action to balance the load of thegiven jobs among active nodes. It also shows that the main source of parallelism is provided by the load balancer, whose role is to split each job into multiple subsets that can be executed on multiple nodes in parallel. After successful execution of threads, each node send its final results back to the sub cloud, then subcloud conquer the results of threads and the output will be passed to the client. In order to reduce the context switch overhead for cloud using dynamic context switch approach an algorithm has been proposed in this research work. Dynamic context switch is defined as the feasible allocation or scheduling of the jobs in such a way that context switches are reduced and so that the waiting time can be reduced. This section discusses the procedure that how dynamic context switch algorithm works.

Fig 2. Dynamic context switch algorithm flowchart Proposed dynamic context switch algorithm is developed considering main characteristics like reliability, performance throughput and resource utilization. However to fulfil these requirements of reducing context switch overheads, in fig | dynamic context switch algorithm is shown.

Table 1 gives the inputs that are given to the simulator. In Table 1 various jobs are given with their burst time. The bur: time means the execution time required for a job to complete. The jobs will run in a parallel fashion in the simulator an context switching will occur when jobs are subjected to run on the simulator. Designed simulator makes all the jobs to get ready for execution in the cloud. Fig 3.giving the detail of which job is going to run and also the other relevant information like burst time, remaining time, status of the job, context switch frequency etc. It allocates the jobs to the cloud in FCFS fashion. When the network is activated it firstly executes the jobs using the classical multiprogramming approach. VII. SIMULATION RESULTS ae eM er Step 5: If remaining time of job is less than equal to 5 sec then no context switch occurs and the same job is executed by the processor, this reduces the context switch and average waiting time is reduced and throughput increases.

Fig 5. Optimal Switching using Dynamic context switch parallel algorithm Fig 5 shows the optimal switching using dynamic context switch parallel algorithm. The simulator first obtains the results by using the classical approach then by activating the optimal switching button on the simulator again the execution of jobs starts using our purposed dynamic context switch parallel algorithm. Fig 5 gives details about the jobs in waiting state, running state and the jobs being executed. It also gives details about various parameters calculated during job execution. Fig 4 showing all the jobs get executed using the classical multiprogramming approach. Fig 4 giving the detail about the _ results obtained using this approach. It gives details about the context switching frequency done to execute all the jobs successfully.

Table 2. Metric’s comparison of different approaches In order to do performance analysis a comparison table has been made for this research work. This section first give th: performance comparison of developed simulator with existing methods and later on comparison of different approache is made using different performance metrics.

Fig 8.Average Context Switch Frequency comparison with existing method. Fig 7.illustrates the graph of Average Waiting (AWT). Therefore it is clearly shown that the purposed method give better results than existing methods as AWT of integrated approach always stay lower than the other existing method lines.

Fig 9.Throughput comparison with existing methods Fig 8.demonstrates the diagram of Average Context Switch Frequency (ACSF) metric. Consequently it is undoubtedly revealed that the purposed technique contributes improved fallouts than prevailing approaches as ACSF in incorporated optimal context switching methodologies line continuously vacation subordinate than the supplementary techniques lines.