Broadcast storm problem—A hidden consequence of content distribution in content delivery networks

Abstract

Content caching at the Internet edge is becoming very popular because of flash crowd problem. Content Delivery Networks (CDN) was evolved for content caching with a complete solution for network reliability, scalability and performance enhancement. A number of researches have been focused on CDN issues like replica server placement, content caching, request routing and accounting management. But still some more issues are yet to be solved. This paper focuses on the concept of Broadcast Storm Problem (BSP) in CDN due to content distribution and request routing. Several approaches are available for content distribution and content caching. When there is any update in any surrogate, the same has to be communicated to all other servers over the network to avoid data inconsistency. Simple flooding or gossiping is generally used for the same, but these approaches are accompanied with BSP. Numerous BSP algorithms have been evolved, but the main concern is the wireless sensor networks. In this paper a comparison among several BSP algorithms has been shown which reveals that counter-based approach is much simpler and can be applied to any network. In addition to the comparative analysis, the counter-based scheme gets modified for CDN.