Task Scheduling using Tabu Search algorithm in Cloud Computing Environment using CloudSim
- 1. 1 Proposal Progress Prepared by Mohd Azarul Ikhwan bin Anuar Salim (BTBL17047526) Supervisor Wan Nor Shuhadah binti Wan Nik Task Scheduling using Tabu Search algorithm in Cloud Computing Environment using CloudSim
- 2. 2 OUTLINE Introduction/Project Background Problem Statement Objective Scope Limitation Literature Review Expected Result Gantt Chart References
- 3. 3 Introduction Today, cloud computing is considered to be one of the hottest trends in the IT industry and the fastest growing buzzword in the computing environment. It is an enticing approach used to architect and remotely manage computing resources. Since it is a “cloud”, the drawback is of course it will not work well under a low speed connections and may be slow depends on the applied scheduling algorithm. Luckily, there is CloudSim which is a simulator that provides simulations of file replication strategies such as replica placement, replication scheduling, and replica consistency maintenance. Tabu search algorithm will be compared to traditional algorithm in term of throughput and turnaround time by using the CloudSim. In conclusion, CloudSim can help researcher in choosing the most efficient algorithm for their computing devices.
- 4. 4 Problem Statement Certain resources scheduling can not adjust well due to big scale of workloads which then lead to starvation. A process had to wait for a resource that is continuously given to other processes. Short processes might need to wait for a long time due to long processes holding the CPU resulting in low throughput. Wrong choice of scheduling algorithm will affect the CPU speed and will waste its resources and time.
- 5. 5 Objective To study Tabu Search algorithm for task scheduling problem in Cloud computing environment. To implement Tabu Search algorithm for task scheduling problem in Cloud computing environment using CloudSim simulator. To analyze the performance of Tabu Search algorithm against one of the traditional scheduling scheme (i.e. FCFS algorithm) in terms of throughput and turnaround time
- 6. 6 Scope CloudSim will be used to apply or implement various scheduling type which user can simulate the real computing performance. By running the simulation, it helps user to choose an algorithm that best suits their computing devices. The end result of the simulation will be analysed and help the user to determine which algorithm that can maximize the throughput and minimized the turnaround time in graph information form.
- 7. 7 Limitation Since it’s a simulation tool ,simulation performance output can not be guaranteed close to that of real time computer performance output.
- 8. 8 Literature Review Author Year Title Purpose Advantages Disadvantages Au, Regina 2006 To Cloud Compute, or Not to Cloud Compute? The popularity of cloud computing is sweeping the entire business world – but there are many pros and cons that need to be considered. • IT Cost Savings • Storage • Capacity • Accessibility • Downtime • Security M.A. Elaziz, S. Xiong, K.P.N. Jayasena 2019 Task scheduling in cloud computing based on hybrid moth search algorithm and differential evolution This paper presents an alternative method for cloud task scheduling problem which aims to minimize makespan that required to schedule a number of tasks on different Virtual Machines (VMs) • MSDE algorithm can efficiently schedule the tasks to the VM while consuming minimum makespan • MSA exploitation ability is not good like its exploration ability REN Xun-yi, WANG Ru-chuan, KONG Qiang 2010 Using Cloudsim to efficiently simulate replica placement strategies This article presents the implementation of a new strategy to improve the CloudSim. • a scale-free algorithm to generate network topology implemented • reduce processing time and storage consumption • Lacks GUI • Typical replica algorithms in Cloudsim cannot simultaneously reduce processing time and storage consumption
- 9. 9 Author Year Title Advantages Disadvantages Monica Gahlawat Priyanka Sharma 2013 Shortest Job First, First Come First Serve and Priority Scheduling Shortest job first and priority scheduling algorithms are beneficial for the real time applications Using very common basic scheduling in comparison Gunho Lee 2012 Modeling a MapReduce job using stochastic values Provider achieve maximum utilization of resources and for user to get application performance requirement with minimum expenditure Difficult to allocate resources in a mutually optimal way due to the lack of information sharing between them
- 10. 10 Methodology In this section, you will find a brief description of how the Cloudsim simulator and Tabu Search algorithm works.
- 11. 11 Users Data Centre Broker Scheduler Cloud Information Services Pool of resources Job Submission Result OutputJob Assignment Figure : Scheduling Model in Cloud Computing Environment Req. for Job Info Resource Info
- 12. 12 The users request for the resources on demand, and the cloud provider is accountable for allocation of required resources to the user to avoid the violation of Service Level Agreement (SLA). The process of Task Scheduling instructs the scheduler to get tasks from the users and asks the cloud information service (CIS) for available resources and their properties. According to the availability of resources and Task Scheduling algorithm, scheduler schedules user submitted jobs on various resources as per requirements. Cloud scheduler is responsible to schedule multiple virtual machines (VMs) to different tasks.
- 13. 13 Figure : Cloudsim class diagram
- 14. 14 The function of the main class in CloudSim : o Cloudlet o Cloudlet Scheduler o Virtual Machines (VM) o Vm Allocation Policy o VmScheduler o Datacenter o DatacenterBroker or Cloud Broker o BwProvisioner o CloudCoordinator o DatacenterCharateristics o Host o NetworkTopology o RamProvisioner o SanStorage o Sensor
- 15. 15 Tabu Search (TS) Algorithm • Method for resolving local search optimization problems • To direct each process to produce the optimal solution, without being trapped in the initial solution found during the continuous process. • To find ways to prevent repetition and find the same solution in an iteration. • Parameters contained in the tabu search method is :- a) Local Search Procedure b) Neighbourhood Structure c) Tabu Condition d) Aspiration Condition e) Termination Criteria
- 16. 16 As you can see the illustration above the gradient leads to the local optimum sometimes. This can be dealt by using Tabu Search Algorithm • We can define prohibited (tabu) states and moves to discourage the search from coming back to previously visited solutions. We have to use a data structure (CircularFifoQueue) to store these prohibited states. This is how tabu search works. Tabu Tenure When a move or state is made tabu it is added to the so-called tabu list with a certain value. This value is the tabu-tenure. The tabu list is usually a queue abstract data type. With each iteration the tabu tenure is decremented by one. When the tabu tenure of a certain move (state) is 0 then this move can be accepted again.
- 19. 19 Expected Result By using tabu search algorithm, it is expected that the throughput of cloud resources can be maximized while minimizing tasks turnaround time as compared to the traditional scheduling algorithm i.e. FCFS algorithm.
- 20. 20 Gantt Chart Activities 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 Topic discussion and Determination 1.1 Discuss title 2 Project Title Proposal 2.1 submit title 2.2 submit brief proposal project 3 Proposal Writing 3.1 Write introduction 4 Write Literature Review 5 Proposal progress presentation and evaluation 6 Discussion and correction proposal and proposed solution methodology 7 Proposed solution methodology 8 Proposed solution methodology(continued) 9 Proof of concept 10 Drafting report of the proposal
- 21. 21 Gantt Chart Activities 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 11 Drafting report of the proposal(continued) 12 Submit draft of report to supervisor 13 Preparation for final presentation 14 Final presentation and Panel’s Evaluation 15 Final report submission and supervisor’s evaluation
- 22. 22 References 1. Au, Regina. (2016). To Cloud Compute, or Not to Cloud Compute?. Innovations in Pharmaceutical Technology. 32-35. 2. M.A. Elaziz, S. Xiong, K.P.N. Jayasena et al., Task scheduling in cloud computing based on hybrid moth search algorithm and differential evolution, Knowledge-Based Systems (2019) 3. REN Xun-yi, WANG Ru-chuan, KONG Qiang et al. Using optorsim to efficiently simulate replica placement strategies. (2010) 4. Chapter 5 Scheduling. Retrieved from https://cseweb.ucsd.edu/classes/sp16/cse120- a/applications/ln/lecture5.html 5. Laguna, Manuel & Barnes, John & Glover, Fred. (1991). Tabu search methods for a single machine scheduling problem. Journal of Intelligent Manufacturing. 2. 63-73. 10.1007/BF01471219. 6. Martins, Simone & Ribeiro, Celso. (2006). Metaheuristics and Applications to Optimization Problems in Telecommunications. 10.1007/978-0-387-30165-5_4.