Problems in the information dissemination of the Internet routing

Abstract

Internet routing is achieved by a set of nodes running distributed algorithms — routing protocols. However, many nodes are resistless to wrong messages or improper operations, unable to detect or correct them. Thus a wrong message or an improper operation can easily sweep almost the whole internet. Such a fragile Internet routing comes from the features of these algorithms and protocols. Besides, the strategies taken by the network equipment manufacturers and administrators also are of important influence. When determining the options or selections in the implementation/operation, they always pay more attention to the expense of a single node or a single area and make some simplifications in implementations and configurations while caring less about the influence on the whole network. This paper tries to illustrate such a scheme is not reasonable at all and suggests the consideration from the view of the overall optimization. From three typical cases involved in the Internet routing, a general model is abstracted, which makes the results significative for more Internet related aspects. This paper evaluates the complexity of the theoretical analysis, then acquires the effect of error information on the whole network through the simulation on the Internet topology. It is shown that even very little error information can incur severe impact on the Internet. And it will take much more efforts of downstream nodes to make remedies. This result is intuitively revealed through the comparisons in the charts and the visual presentations. Then a hierarchical solution to establish the upgrade plan is given, which helps to upgrade the nodes of the network in a most efficient and economical way.