Cloud storage with key-value stores over content-centric networking architecture

Abstract

Recently, cloud storage systems have been used for various purposes. For example, they are used for content delivery because it is possible to improve responses distributing data for end users by geographically distributed multiple storage nodes. Applications can use them as infrastructures for computing. In the cases, it is desirable that each client node can retrieve the desired data from a nearby storage node for the faster response for end users. In Internet-scaled cloud storage systems, Key-Value Stores(KVS) is used to manage data and it provides REST(Representational State Transfer) API, in which each data has a unique URI(Uniform Resource Identifier) and clients can access the data by the defined methods. The architecture has a high affinity to the CCN (Content-Centric Networking) architecture, on which each data has a unique “Name” to identify it. Over the CCN architecture, data is cached in intermediate-nodes relaying it and the communication is not based on server's IP address but data's “Name.” Therefore each client node retrieves desired data from a nearby node caching it. In this paper, in order to improve the response for end users, we propose the cloud storage with KVS over the CCN architecture as a new cloud system. The response time of proposed system is measured with a simple simulation. The results show that the proposed system has better performance than the cloud storage built in distributed data centers over the TCP/IP architecture.