Article Preview
Top1. Introduction
Cloud computing arose as a result of the massive expansion in internet data processing. It is crucial in providing computing services over the internet. Cloud computing provides cloud users with data storage, computational power, and other computer system resources without requiring direct active control. Cloud providers provide leverage virtualization technologies with virtual machines by Stuti and Prashant (2014), to cloud users with computing power. Effective task/job scheduling is necessary to enhance the QoS in cloud system as numerous users demand cloud services. Optimal allocation of resources allows for efficient job scheduling in a finite amount of time in order to achieve the specified level of Quality of Service (QoS) as given by Vinothinna, V. and Rajagopal, S. (2022). Efficient task/job scheduling allows optimal allocation of resources among the requested tasks in a finite time to achieve desired quality of service (QoS). To enhance quality of service, optimal scheduling of tasks to appropriate resources is required in cloud. So, the important element of cloud infrastructure is efficient scheduling algorithm to schedule the task or resource. The most forms of scheduling algorithms are heuristics scheduling, static scheduling, dynamic scheduling, unit cloud service scheduling and work flow scheduling which is described by Amanpreet and Bikrampal (2022). Although, scheduling techniques are often classified basis on the complexity of scheduling algorithm such as heuristic job scheduling, meta-heuristic job scheduling and hybrid job scheduling explain by Mohit and S.C. Sharma (2019), Tyagi, R., Gupta, S.K. (2018), as well as Tripathi G., and Kumar R. (2022). There are several heuristic algorithms in cloud such as LJFP, SJFP, Min-Min, MET, Max-Min, MCT etc. All the heuristic algorithms are based on static and non-optimized scheduling algorithms. Metaheuristic scheduling is based on dynamic scheduling. Examples of metaheuristic scheduling in cloud computing are PSO, ACO, DSOS, GSA, HA etc. Hybrid metaheuristic scheduling algorithms are used to find improved performance in cloud computing. Examples of hybrid metaheuristic job scheduling are PSO and Q-learning, PSO and firefly, firefly and SA (simulated annealing), hybridized whale optimization algorithm. User and CSP are two important entities in cloud system. Users that use the cloud want their programs to run quickly and at a low cost.
A Harmony search based meta-heuristic scheduling algorithm named as New HS(New Harmony Search) has proposed to optimize some QoS parameters like throughput, execution cost, makespan time, execution time and task rejection ratio etc. In a cloud, there are many data centers available to execute requests, but due to reduced latency, requests are directed to the nearest data center. If requests are not routed to the adjacent data centers in cloud, there is a risk of high latency, which can influence some quality of service (QoS) criteria such as deadlines and response times. After changing the QoS parameter, the number of SLA violations increases. In a data center, a task request handler or gatekeeper receives user service requests. The Turing test is used by the job request handler to determine if the coming request is from an attacker or a legitimate user. If the request comes from an attacker, obstruct the user based port address, the source IP address and other factors.