Distributed multi cast caching technique

Abstract

A caching arrangement for the content of multicast transmission across a data network utilizes a first cache which receives content from one or more content providers. Using the REMADE protocol, the first cache constructs a group directory. The first cache forms the root of a multilevel hierarchical tree. In accordance with configuration parameters, the first cache transmits the group directory to a plurality of subsidiary caches. The subsidiary caches may reorganize the group directory, and relay it to a lower level of subsidiary caches. The process is recursive, until a multicast group of end-user clients is reached. Requests for content by the end-user clients are received by the lowest level cache, and forwarded as necessary to higher levels in the hierarchy. The content is then returned to the requesters. Various levels of caches retain the group directory and content according to configuration options, which can be adaptive to changing conditions such as demand, loading, and the like. The behavior of the caches may optionally be modified by the policies of the content providers.

Description

technical field [0001] The present invention relates to the multicasting of information over a data network. In particular, the present invention relates to an improved caching technique for multicast over data networks. Background technique [0002] In recent years, the push model of data distribution over the Internet and other client-server networks has grown in popularity. In modern versions of this model, a server "multicasts" data to a subset of interested clients on the network, called a "multicast group". Any interested user can become a so-called "listener" by joining the group. [0003] Because of this feature, the application of the propulsion model is closer to the broadcasting paradigm of radio and television than to the interactive paradigm of the World Wide Web. Thus, a broadband network such as cable TV or satellite can be used as a very efficient medium for "pushing data". Unfortunately, these networks are currently only one-way. That is, data such as T...

AI Technical Summary

Problems solved by technology

Although there are some optimizations, receiver acknowledgment via ACK / NAK is still required, and disadvantageously, the data needs to be resent to the selected receiver based on the information about the NAK, because a large number of receivers send ACK / NAK can still cause network congestion

Images