Facilitated by high-speed internet connectivity, the adoption of cloud computing has surged dramatically due to its distinctive attributes such as pay-as-you-go billing, ubiquitous accessibility, and minimal support requirements in the realm of service and resource delivery. While cloud computing draws inspiration from distributed, parallel, and grid computing paradigms, it also inherits their associated complexities and concerns. Among these challenges, optimizing performance stands out prominently. Enhancing performance can be achieved through the development of efficient load balancing algorithms. Efficient algorithms satisfy users and make optimal uses of resources of the cloud and hence improve performance. Several load balancing algorithms were designed to respond to performance problems. In this chapter, three well-established load balancing algorithms, namely round robin, equally spread current execution, and throttled, underwent a comprehensive analysis. This examination sought to unveil both the strengths and weaknesses inherent to these algorithms.
TopIntroduction
Cloud computing is akin to the electricity grid, with computing resources connected centrally rather than each entity generating their own (Kashyap & Viradiya, 2014). This arrangement offers advantages in terms of cost-efficiency and time savings. Modern cloud computing shares similarities with traditional mainframe computing but distinguishes itself through its remarkable advancements in speed, storage capacity, memory capabilities, and cost-effectiveness. Despite these advantages, direct access to cloud resources introduces security concerns, representing a primary challenge in cloud computing (Kashyap & Viradiya, 2014). Furthermore, performance optimization and load balancing pose additional hurdles in ensuring equitable distribution of millions of user requests to data centers while maximizing resource utilization and user satisfaction (Qi & Boutaba, 2010). Researchers are actively addressing these issues by devising frameworks and strategies, with particular emphasis on mitigating vulnerabilities that may expose vulnerabilities to potential attackers. The cloud continues to experience sustained growth within the business community, primarily due to its ability to offer cost-effective solutions to customers. This expansion prompts cloud service providers to reevaluate their resource management strategies. Efficient resource management is crucial in accommodating the increasing number of customer requests without the necessity of expanding physical infrastructure. Additionally, there arises a need for equitable distribution of customer requests across available resources to ensure optimal utilization. In the realm of Cloud computing, providers are dedicated to enhancing overall performance. Various strategies are employed to achieve this goal, and among them, load balancing stands out as a pivotal technique. Load balancing has garnered significant attention in recent years within the Cloud computing landscape (Reddy & Reddy, 2014). Its importance cannot be overstated, as it brings a multitude of benefits, including improved revenue, heightened user satisfaction, enhanced performance characterized by reduced response times, increased acceptance rates, and more efficient resource utilization. But to obtain all these services providers should use efficient virtual machine (VM) load balancing policy.
In the contemporary cloud computing environment, a diverse range of VM load balancing policies has gained prominence. Among these, the Throttled load balancing algorithm, Equally Spread Current Execution, and Round-robin are widely recognized. These algorithms play a crucial role in optimizing resource allocation and ensuring efficient performance within cloud-based systems.
Operating in a circular fashion, the Round-Robin load balancing algorithm distributes tasks evenly among virtual machines. When a request enters the data center, it randomly selects the first available virtual machine to handle the assignment. Subsequent requests are then allocated to virtual machines in a cyclical manner. Notably, this algorithm efficiently utilizes available resources without considering the current status of the virtual machines, such as whether they are busy, available, or their current load. Consequently, this approach evenly distributes requests across all virtual machines but may result in longer wait times for most requests before execution, in addition to their actual execution time. This leads to a higher overall response time for requests, making the Round-Robin algorithm susceptible to having a maximum response time (T. S. S. K. S. M. & Ramasubbareddy, 2019).