A peer to peer resource discovery scheme for provisioning in cloud using multi attribute range query
Rayhanur Rahman, Nadia Nahar, Rayhanul Islam, Kazi Muheymin Sakib · 2014
Provisioning in cloud functions through obtaining information of physical and virtual machines, known as resource discovery. The existing such resource discovery schemes are centralized in nature, imposing scalability and single point failure issues in overall datacenter operations. As peer to peer (P2P) approaches are inherently scalable and invincible to single point vulnerability, this paper proposes a P2P based decentralized and scalable resource discovery method for provisioning in cloud. By using P2P architecture, routing data tuples and publishing those by storing inside nodes are also presented here. The paper also proposes searching appropriate multi attribute provisioning data stored in the nodes and fail-safe policies in case of physical machine failure. For simulation purpose, a datacenter environment was developed using a P2P overlay network over which provisioning functioned based on the proposed resource discovery scheme. The experiment showed that the proposed approach is 46.15% faster on average in case of multi dimensional range querying. It also proved to be scalable with increased number of virtual machines in resource publishing with a slight penalty of non uniform distribution of resources in the datacenter.