Dynamic Workload Handling & Performance Analysis of Switches in Cloud Storage System

Cloud computing is a key area of today’s IT industry; it allows a huge number of clients to access a wide range of resources through internet technology. Cloud storage offers a low-cost, user-friendly service to clients. The cloud network backbone is very complex; it’s a collection of various devices to form a cloud environment. Due to their complex architecture, they need various types of management services, such as congestion controllers, load balancers, service request schedulers, and efficient routing techniques. In the traditional approach, cloud storage uses the round-robin green Heros scheduler, which schedules based on the service provided by the cloud server but it needs to improvise with new features of recent machine learning technology, which fairly distributes the load efficiently with a low network resource cost-based method. This article investigates existing systems, their load handling capability, and comparative analysis with the proposed dynamic workload distribution for cloud storage (DWCS). The proposed dynamic load distribution system implements three-tier cloud architecture. In this architecture, the core, aggregator, and access switch are involved in the routing process, which efficiently handles requests from users and accesses cloud storage effectively. The experimental outcome of DWCS compares with the traditional scheduling techniques of round robin, Heros and green. The result shows that DWCS power consumption is minimized, and it's more capable of handling load in each layer of switches compared to existing scheduling techniques for cloud data centers.