(IJACSA) International Journal of Advanced Computer Science and Applications, Vol. 6, No. 2, 2015 Resource Provisioning in Single Tier and Multi-Tier Cloud Computing: ―State-of-the-Art‖ Marwah Hashim Eawna Salma Hamdy Mohammed El-Sayed M. El-Horbaty Faculty of Computer and Faculty of Computer and Faculty of Computer and Information Information Information Sciences Sciences Sciences Ain Shams University Ain Shams University Ain Shams University Cairo, Egypt Cairo, Egypt Cairo, Egypt Abstract—Cloud computing is a new computation trend for 2) Dynamic Resource Provisioning: the basic delivering information as long as an electronic device needs to fundamental idea in the latter way is providing the resources access of a web server. One of the major pitfalls in cloud computing is related to optimizing the resource provisioning and based on the application needs, this helps the provider to allocation. Because of the uniqueness of the model, resource assign the Non-loaded resources (which become free to used provisioning is performed with the objective of minimizing time now) to the new users. This method reduces a fraction of and the costs associated with it. This paper reviews the state-of- providers’ development costs by utilizing current available the-art of managing resources of the cloud environments in resources and beside that the user can happily just pay for the theoretical research. This study discusses the performance and analysis for well-known cloud provisioning resources techniques, amount of the resources which were really used [3]. single tier and multi-tier. Moreover, the parameters of resource provisioning is Keywords—Cloud Computing; Resource Provisioning presented as: 1) Response time: The resource provisioning algorithm I. INTRODUCTION designed must take minimal time to respond when executing The cloud environment describes a company, organization the task. or individual that uses a web-based application for every task 2) Minimize Cost: From the Cloud user point of view cost rather than installing software or storing data on a computer. should be minimized. All cloud environments are not common but a move toward 3) Revenue Maximization: This is to be achieved from the this is a long-term goal for cloud computing enthusiasts and Cloud Service Provider’s view. cloud capitalists. 4) Fault tolerant: The algorithm should continue to Many challenges are influenced to adopt the cloud provide service in spite of failure of nodes. computing technology such as security, resources allocation, 5) Reduced SLA Violation: The algorithm designed must resources provisioning and others. In provisioning resource for be able to reduce SLA violation. cloud computing environment the major challenge is to 6) Reduced Power Consumption: virtual machine determine the right amount of resources required for the placement & migration techniques must lower power execution of work in order to minimize the financial cost from consumption [4]. the perspective of users and to maximize the resource The rest of this paper is organized as follows: Sections 2 utilization from the perspective of service providers [1]. contains a review of resource provisioning in single tier and The resource provisioning must meet Quality of Service multi-tier cloud environments. Section 3 illustrates the most (QoS) parameters like availability, throughput, response time, popular efficiency of Single tier and Multi-tier architectures. security, reliability etc., and thereby avoiding Service Level Finally, Conclusions and future works are given in Section 4. Agreement (SLA) violation [2]. There are entirely two generic II. RELATED WORK way of resource provisioning, Static and dynamic. This section explains two basic architectures are dependent 1) Static Resource Provisioning: usually provides the in the resources provision in cloud computing environment peak time needed resource all the time for the application. In like single tier and multi-tier architecture. this kind of provisioning most of the time the waste of resource A. Single-tier technique due to workload is not in a peak, but resource providers provide the maximum required resource to prevent SLA A single-tier architecture of cloud computing has a set of violation. servers that are used to provide resources by receiving 213 | P a g e www.ijacsa.thesai.org (IJACSA) International Journal of Advanced Computer Science and Applications, Vol. 6, No. 2, 2015 requests from user in presentation server and looking Xiaotang Wen el al. [9] improved algorithm to provide information in application server and store the information in resource by combine Ant Colony Optimization (ACO) with database server. particle swarm optimization algorithm to improve the efficiency of resource scheduling in cloud computing environments. Table 1 summarizes the advantages and disadvantages of single tier mechanisms are surveyed in this section. TABLE I. RESOURCE PROVISIONING ALGORITHMS IN SINGLE TIER TECHNIQUES Parameters Techniques Attributes Authors Scheduling minimize Tarun model based on the make span goyal et al. [5] GA SA-based minimize Monir approach for execution time Abdullah et sche-duling in al. [6] cloud envi- ronment PSO reduce the Shaobin Fig. 1. Single tier architecture in cloud computing algorithm in task average Zhan et al. [7] Time resources running time, As show in figure1, single tier architecture there are no scheduling and raises the interactive between the servers and is not allowed to share strategy rate availability of resources between server that cause consumption of resources and loss of time when compared with multi-tier architecture. resources Use PSO and less Talwinde Some of the existing techniques involving nature-inspired SA algorithm execution time r et al.[8] meta-heuristics have become the new focus of resource as compare with allocation. For example, Tarun goyal et al. [5] Scheduling a existing model based on minimum network delay using Suffrage algorithm Heuristic coupled with Genetic algorithms for scheduling sets Use PSO and Efficiently Xiaotang of independent jobs algorithm is proposed, the result show that ACO algorithm and speed to et al.[9] the schedule of multiple jobs on multiple machines in an provide resource efficient manner such that the jobs take the minimum time for completion. B. Multi-tier technique This architecture partitions the application process into Othman et al. [6] proposed a novel Simulated Annealing multiple tiers. Each tier provides certain functionality. The (SA) algorithm for scheduling tasks in cloud environments. benefit of such architecture is that it can provide a high level SA exploits an analogy between the way in which a metal of scalability and reliability. However, the resource allocation cools and freezes into a minimum energy crystalline structure among these tiers will be more difficult due to the (the annealing process) and the search for a minimum in a interdependency between the tiers. A multi-tier cloud more general system. Results show that this approach for job computing application may span multiple nodes. Specifically, scheduling not only guarantees the QoS requirement of most multi-tier cloud computing applications use 3-tier customer job but also ensures to make best profit of cloud architecture. providers. It has also concern about real execution time of jobs in different systems as well as deadline and penalty cost in the As shown in figure 2, the first tier, named presentation tier, algorithm. consists of Web servers. It displays what is presented to the user on the client side within their Web browsers. For the Web Shaobin Zhan et al. [7] introduces an improved Particle server tier, it mainly has three functions: Swarm Optimization (IPSO) algorithm in resources scheduling strategy by Add simulated annealing into every 1) Admitting/denying requests from clients and services iteration of PSO, through experiments, the results show that static Web requests. this method can reduce the task average running time, and 2) Passing requests to the Application server. raises the rate availability of resources. 3) Receiving response from Application server and Talwinder Kaur et al. [8] Improved Particle Swarm sending them back to clients. Examples of Web servers include Optimization (IPSO), Simulated Annealing (SA) Algorithm, Apache Server and Microsoft Internet Information Server and Hybrid Particle Swarm Optimization-Simulated (IIS). Annealing algorithms based on utilizing and scheduling The second tier, named business tier, consists of resources. The experiment show that by using this algorithm Application servers. Business logic processing is performed at provisioning resource in very less time as compared to the this tier. There are also three functions at the Application existing algorithm. server tier: 214 | P a g e www.ijacsa.thesai.org (IJACSA) International Journal of Advanced Computer Science and Applications, Vol. 6, No. 2, 2015 1) Receiving requests from the Web server. TABLE II. RESOURCE PROVISIONING ALGORITHMS IN MULTI-TIER 2) Looking up information in the database and processing TECHNIQUES the information. 3) Passing the processed information back to the Web Parameters of Techniques Attributes References server. novel double the Bhuvan The last tier, named data tier, consists of database servers. dynamic application Urgaonkar It handles database processing and data accessing. Database provisioning capacity et al. [11] server tier is used to store and retrieve a Web site’s Time technique reduced the overhead of information (e.g., user