A Hybrid Approach for Task Scheduling in the Cloud Environment

Introduction

Basically, distributed computing focused on the classical framework which works on the model of on-request and it is very advantageous service and it organize the resource pool and it can compute and manage the resources like servers, storage and other application etc. The other purpose of distributed computing is to quickly provide the administrative and various service provider interactions. In other word, we can say that when we are talking about cloud computing, here cloud means a service provider that provides various services and resources to all the users. Cloud computing is a distributed computing model which are interconnected to various users and it virtualize many personal computers that are connected with each other. The purpose is to provide at least one consolidated algorithm for the various resources which manages the various tasks without violating the service level agreement. The SLA is between the various consumers and the service providers. There are few difficulties with distributed computing like security issues, execution issues, reliability and so on. Distributed computing helps in spreading the computational workload among various other servers. So, the distributed computing is an execution environment where the various other resource can meet with the requirements of other applications. The requirement includes:

• 1.

  Security: Security is the most important requirement in the data and processing of data. Security remains with the data centers of various enterprises which one public cloud is attached the sharing of data and resources.

• 2.

  Location: Location is the most important requirement because it is used with some of the applications for make it more responsive and for making the performance of the application better.

• 3.

  Redundancy: The next requirement in cloud computing is redundancy. Redundancy helps to mitigate the various other applications and other large-scale enterprises.

• 4.

  Regulations: The most important requirement of distributed cloud is the regulations. So that data cannot leave the geographical area. It is shared only on the location where it is permitted to.

Hence, the service providers must ensure for the end-to-end management for the placement of data, interconnections and various other computing processes and all these must appear as a single solution from cloud computing point of view. Here, Content Delivery Network (CDN) is the most common example of distributed cloud environment where storage is done based on the geographical diverse regions. The purpose is to reduce the latency of delivery to various resources.

The first very common resource issue is the managing of various tasks. In cloud computing, task scheduling refers to allocation of various task to virtual machines by physical machines to increase the processing of tasks and it will also increase the utilization of resources. So, to get the better performance from the cloud, task scheduling plays a vital role. There are numerous tasks that might get scheduled to increase the performance of the system and for minimizing migration. Thus, for this work task management plays vital role to have improved reliability and flexibility of the system in cloud computing. Most of the algorithms are rule based to perform the implementation but rule-based task scheduling performed not very well to solve the complex problems of task scheduling. Another point in task scheduling is the allocation and scheduling of resources related to Quality of Service (QOS) which will affect the cloud computing service providers. So, another important issue here is the resource scheduling, which is a critical issue in cloud-based environment. Distributed computing must need to scale with the various numbers of assignments and various customers and must have the scheduling algorithm and that can spread various assets. They form key point in the investigation and must follow the different methods of probability. Examples are Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO) and so on.

In this paper, author proposed a model of task scheduling which is based on various objectives and use them with combination of harmony search optimization and cuckoo search algorithm. On the basis of these multi objectives functions, one can have obtained the task scheduling from the set and also it is used for the memory usage, energy consumption and most importantly the cost. Following contributions are made in the research for the process of task scheduling: