In order to attain high performance computing (HPC) and utilization of computing resources, the elementary concept of the distributed system is utilized to perform cluster, grid and cloud computing. The distributed application determines the distributed paradigms. Virtual machines (VMs) can comprise of separately functioning operating systems and applications. Virtualization is the core concept of resource pool and management. Hypervisor assists in attaining hardware virtualization and is segregated into Type 1 and Type 2. The Type 1 hypervisor is the bare-metal hypervisor that is installed directly on the x86-based hardware and renders direct access to the hardware resources . The Type 2 hypervisor is the hosted hypervisor that is installed and run as an application on an operating system. The operating system and the applications of a VM function autonomously without any mutual interference. VM is migrated without downtime and VM failure does not affect the distribution of resources among VMs. The service level agreement (SLA) and the quality of service (QoS) must be managed by the load balance (LB) policy. The main causes of SLA violation are scattered data among heterogeneous servers, hot spot, load imbalance, and weak resource management (Zhang & Zhang, 2010). The occurrence of the load imbalance is when the demands in the heterogeneous environments are frequently changing. Load imbalance can be managed through LB between high load and low load machines. The management of LB is difficult on high resource demands that change frequently. The factors that help in the management of LB are information policies, location to migrate VM, selection of VM, and transfer of load.
Load Balancing
The management of demanded resources is known as load balancing mechanism in which the workload is distributed among the VMs. The mechanism of load balancing is a vital component of the hypervisor, which dynamically or statically manage the load imbalance in a distributed manner on the available VMs. The CPU, memory and network components are virtualized to maximize the utilization of resources. In computing, the distribution of workloads across multiple computing resources, such as disk drives, CPUs, network links, computer clusters, or computers, is enhanced by load balancing (Sreenivas et al., 2014). Figure 1 shows the load imbalance scenario in distributed computing. The figure shows that load imbalance is a cause of system inefficiency. The aims of load balancing are to avoid the overload situation in nodes, minimize the migration and response time, maximize throughput, and optimize the resource utilization. Availability and reliability may be increased through redundancy by utilizing multiple (and not single) components with load balancing. Dedicated software or hardware, such as a domain name system server process or a multilayer switch, is generally involved in load balancing.
In distributed computing, Load Balancing is vital for capable operations in appropriated circumstances. As Cloud computing is increasing and consumers are asking for more organizations and better results, load altering for the Cloud has transformed into an especially captivating and basic examining range. Load altering enables all processors in the structure or each of the center points in the framework to generally do the proportionate measure of work at any given point of time. It is utilized for attaining a better resource utilization ratio and service provisioning, thereby enhancing the systems’ general performance. Incoming tasks from different locations are received by the load balancer and then distributed to the data center, for a proper load distribution. Our research objective maximizes the CPU utilization of VMs with efficient results in load balancing metrics (Alakeel, A. M., 2010) Figure 2 shows the role of load balance manager.