Restoration methods for cloud multicast virtual networks. (15th January 2017)

• Record Type: Journal Article
• Authors: Ayoubi, Sara; Assi, Chadi; Chen, Yiheng; Khalifa, Tarek; Shaban, Khaled Bashir
• Is Part Of: Journal of network and computer applications. Volume 78(2017)
• Page Start: 180
• Page End: 190
• Publication Date: 2017-01-15
• Subjects: Cloud computing -- Virtualization -- Multicast -- Survivability; Microcomputers -- Periodicals; Computer networks -- Periodicals; Application software -- Periodicals; Micro-ordinateurs -- Périodiques; Réseaux d'ordinateurs -- Périodiques; Logiciels d'application -- Périodiques; Application software; Computer networks; Microcomputers; Periodicals; 004.05; 004
• DOI: 10.1016/j.jnca.2016.11.009
• Languages: English
• ISSNs: 1084-8045

Abstract

Abstract: Providing reliability guarantees for services hosted in cloud data center networks is a well studied problem that received much attention from the literature in recent years. Yet, the existing literature failed to account for the mode of communication that these hosted services may exhibit. Rather, it is assumed that the constituent Virtual Machines (VMs) are always exchanging traffic in a one-to-one communication mode. This paper is concerned with the problem of restoring failed services with a one-to-many communication mode, the mode of communication for a multitude of multicast applications and services hosted today in the cloud. Through motivational examples we show that the problem of restoring multicast Virtual Networks (VNs) greatly differs from that of unicast services, thereby demanding separate attention. To this extent, we formally define the problem of restoring multicast services in event of facility node failure, and we mathematically formulate it using an Integer Linear Programming (ILP) model. Further, we propose REAL: a multicast virtual networkre storational gorithm. REAL distinguishes the failure of multicast source node from that of multicast terminal nodes, and adopts a hop-to-hop search to handle receiver or terminal node failures, and a path convergence approach to recover from source node failures. We evaluate our proposed schemes against a Greedy and a Steiner-based restoration schemes, and we show that our suggested method outperforms its peers in terms of restoration ratio and total achievable revenue.