Article Preview
TopIntroduction
The cloud computing model consists of several DCs, which might be located in one or multiple geographical regions. These DCs are to provide different computing resources to the users and achieve their satisfaction (Naha, 2016). The transitions to a SaaS have been increased quickly and rejecting requests might happen when many users are trying to use the service at the same time, especially in a specific period which is considered an important issue (Shi, 2014). On the other hand, allocating the optimal DC is based on the concept of “Best Data Center” to process a particular user’s request. This allocation depends on the priority of the user who submitted the request (Arya, 2017).
The most common concern faced by the Internet web provider is how to resolve the congestion problem that is caused by the large number of user requests at specific times. The main issue of performance in the cloud is how to serve the user requests and balance them in an efficient way. Load Balancer (LB) and Cloud Service Broker (CSB) play the main role in controlling the traffic using different policies. Therefore, these techniques are receiving more attention from researchers, and designing these approaches and techniques to distribute workload fairly and efficiently will improve user satisfaction and achieve a high level of performance for the whole cloud system.
CSB is an entity that acts as an intermediary between cloud clients and cloud service providers and selects the appropriate DC for each user request. CSB uses the existing policies to find and route user requests to the best DC that would satisfy user service requirements. In other words, a cloud services broker intermediaries cloud provider with consumers and make it possible for companies to choose cloud services and offerings that are appropriate to their requirements. According to NIST (National Institute of Standards and Technology), “Cloud service broker is an entity that manages the use, performance, and delivery of cloud services and negotiates relationships between cloud providers and cloud consumers” (Khurana, 2017).
Several service broker policies use a few factors for selecting the best data center. These factors are the distance between data centers and users, response time, and current execution load (Manasrah, 2017). The cost of the data center usage is an important factor that should be considered as well. Also, using a single-entry point to receive incoming requests is a bottleneck (a single point of failure). Also, it causes a delay in serving users or rejecting them. Therefore, the main objective of this paper is to balance the user workload that comes from different geographic allocations in an efficient way to improve performance. This kind of balancing is for enhancing service broker policy that selects the optimal DC considering several factors together which are response time, processing time, DC load, and DC cost. Also, to ensure efficient distributing of user requests in optimal execution time (i.e., minimized response and reducing the rejected requests).
In this paper, a new model to address user requests and workload concerns at the main entry point is proposed and presented. Also, it shows how to select the optimal DC to serve the user’s requests. The main contributions of this paper include the followings:
- •
Designing a model for cloud service broker by using multiple entry points for maximizing receiving number of user requests.
- •
Developing the policy for cloud service brokers by considering several factors that are added to the developed policy to improve the performance.
- •
Recent statistics about internet users and the variation of user workload are presented.
- •
A comparative study of the popular open-source simulation tools is explained.