Task Scheduling in Cloud Computing

Project Domain

Cloud Computing

Introduction

Cloud computing is a core technology offering computing resources on a pay-per-use basis. Scheduling the incoming workload by cloud providers is a challenging task to satisfy numerous quality of service requirements. This project aims to evaluate the performance of basic scheduling algorithms on a publicly available dataset, GoCJ: Google Cloud Jobs Dataset.

Functional requirements:

        Develop a datacenter

Make a datacenter with 10 Virtual Machines (VM) according to the specifications given in Table-5 in https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10431793

        Load tasks from the GoCJ dataset

Read the length of each task from the GoCJ dataset for various number of tasks. The dataset is available at https://data.mendeley.com/datasets/b7bp6xhrcd/1

        Implementing scheduling algorithms

Implement the following algorithms to schedule the loaded tasks on available VMs:

        First Come First Served (FCFS)

        Shortest Job First (SJF)

        Earliest Finish Time First (EFT)

        Evaluation parameters

Evaluate the performance of all algorithms based on makespan, resource utilization, and throughput.

        User Interface

Develop a graphical user interface to run and display the results of scheduling various number (100-1000) of tasks.

        Reporting

Save the results computed in step-4 for detailed analysis.

        Documentation:

Document the system design, scheduling algorithms, experimental setup, and evaluation results.

Tools and language:

Languages: Python

        IDE: No restriction Supervisor:

Name: Dr. Zulfiqar Ali Khan

Email ID: zulfiqarmrd@vu.edu.pk

MS Teams ID: zulfiqar.vu@outlook.com