Future Request Predicting Disk Scheduler Computer Science
A Virtual machine monitor (VMM) allows multiple virtual platforms to share the same physical machine safely and fairly. VMM provides separation among the virtual machines and manages their access to hardware resources .
The disk scheduler within the VMM plays a very important role in determining the overall fairness and performance characteristics of the virtualized system.
The disk is a resource which is shared by different applications. Thus, there must be some scheduling criteria to use the disk properly. Therefore, Disk scheduling can be defined as a method of selecting threads or processes from the request queue and scheduling the ready processes.
There occurs advancement in the disk technology. Virtualization has brought several challenges to disk scheduling. Nowadays, disk schedulers designed for Operating system are based on the latency characteristics of the underlying disk technology. In the virtualized system the disk latency characteristics are found to be different than the traditionally assumed characteristics. Disk latency characteristics of the guest operating system depends on the underlying disk as well as the additional queuing and processing that happens in the virtualization layer. Furthermore, the virtual machine has been provided with the virtual disk shows the limitations of existing disk schedulers . Thus, there is a need to re-examine the disk scheduler for the virtual environment.
Till date the contribution on disk scheduling in virtualized environment is comparatively less. The paper aims to study whether the traditional disk scheduling algorithm still provides performance benefits in the layered system consisting of virtualized operating system and underlying virtual machine monitor.
In addition, performance provided by the different scheduling algorithm depends on the workload characteristics. There are various applications in the system that needs to be executed. There are multiple requests to be handled by a single application. Thus, the system performs such that it takes a request from the application and sends it to the disk, thus executing the request sequentially producing the response. It optimizes the performance of various requests. The operating system can improve the overall system performance by keeping the disk as busy as possible.
The scheduler in the virtualized environment has a significant role in improving the overall fairness and performance characteristics of the system .Therefore, there is a need for the implementation of the recently developed disk scheduling algorithms in virtualized system and to see which one is better than the existing ones.
In any operating system the aim to perform disk scheduling is to increase the throughput and to decrease the seek time and rotational latency and keep disk as busy as possible. Even in the virtualized environment the disk scheduling is performed for the same reason. To perform the disk scheduling various scheduling algorithms are used.
Â© 2012 JCSE
www.Journalcse.co.ukThe most important issue while performing disk scheduling is to consider two important parameters, the throughput and fairness. Throughput refers to the number of requests that are completed in some period. A totally fair system is the one which ensures that the mean response time of the disk requests is the same for all processes. In order to improve the system performance the best way is to modify existing disk scheduler . Providing system performance guarantees leads to performance isolation that means the performance experienced by an application should not suffer due to variations in the workload from other applications .
Throughput and rotational latency are the key parameters of disk scheduling technique. Traditional disk scheduling algorithms which are designed for OS aims to improve the disk throughput and performance of the system.
In this paper, the existing disk scheduling algorithms like High Throughput Token Bucket algorithm which is designed for traditional OS are used inside the virtual operating system. These algorithms are non-work conserving i.e. they predict the future request arrival. In this paper, it is studied whether the concept of predicting future requests still improve the throughput and performance of the virtualized system.2
Article name: Future Request Predicting Disk Scheduler Computer Science essay, research paper, dissertation