Informatics Publishing Limited

Laouratou Diallo ¹, Aisha-Hassan A. Hashim ¹, Rashidah Funke Olanrewaju ¹, Shayla Islam ¹, Abdullah Ahmad Zarir ¹

Affiliations
1 Faculty of Engineering, International Islamic University Malaysia, Jalan Gombak - 53100, Kuala Lumpur, MY

Abstract

Cloud computing is the latest and the most used type of distributed computing systems and also it covers most of their features. It has been widely used for its enormous benefits and its ability to cope with large scale data such as workflows and big data applications. On the other hand, scheduling algorithms; starting from traditional to Hyper-heuristic; are widely used in computing systems such as cloud computing to monitor the use of resources. However, these scheduling algorithms vary in term of their performance and most of these traditional and simple scheduling algorithms may not be efficient for large scale data. Although many scheduling algorithms have been implemented for cloud computing, it has been realized that most of the applications nowadays require different objectives that simple scheduling algorithms fail to achieve. Either one of the objective is violated or the results are far from the optimal solution. In this direction, this paper first gives review of some previous scheduling algorithms used in cloud. Then, it proposes a type of swarm intelligence called Particle Swarm Optimization (PSO) algorithm to diminish cost though meeting deadlines. The proposed method is evaluated using CloudSim and big data applications are used as sample of applications. From the results, it can be seen that PSO works better for big data applications and the cost is reduced to more than half when compared with ordinary scheduling algorithms such as First-Come-First-Serve (FCFS).

Keywords

Cloud Computing, Hadoop and Big Data, Scheduling, Swarm Optimization.

Full Text

   

Azana Hafizah Mohd Aman ¹, Aisha-Hassan A. Hashim ¹, Azween Abdullah ², Huda Adibah Mohd Ramli ¹, Shayla Islam ¹

Affiliations
1 Kulliyyah of Engineering, International Islamic University Malaysia, Jln Gombak 53100, Kuala Lumpur, MY
2 SOCIT, Taylor’s University, Jalan Taylors, Subang Jaya 47500, Selangor, MY

Abstract

Background/Objectives: Mobile data traffic over IP has grown very rapidly in size. This huge increase in size creates high demand performance on network that supports mobile multicast services. This motivates the development of a better performance procedure with better signaling cost. Methods/Statistical analysis: A mathematical evaluation of signaling cost for multicast network mobility management namely Proxy Mobile IPv6 (PMIPv6). The signaling cost is derived from a novel combination of Multicast only Fast Reroute (MFR) and predictive Context Transfer (CT) with network mobility management. The signaling cost is designated base on the improved signaling call flow of the advanced method combination. It is calculated as the location updates and the packet delivery cost of the call flow. Findings: This combined procedure helps to mitigate unnecessary multicast network mobility traffic as usage increases. From the results it is clearly shows that the location update for the advanced signaling cost remain consistent regardless of the traffic usage. Where else for the standard method the signaling cost increases significantly, in parallel with the traffic increase. This is hardly because of the reduced location updates and the packet delivery cost of the novel combination procedure. Therefore through this implementation better signaling cost formula is brought forward. Application/Improvements: The advanced signaling cost helps to overcome performance degradation in multicast mobility management application such as live Internet video, Internet-video-to-TV, online video, webcam viewing, video conference and web-based video monitoring.

Keywords

Fast Reroute, Mobility Management, Multicast, Signaling Cost.

Full Text