The Internet consists of nearly 12,000 autonomous systems (ASes)
that exchange routing information using the Border Gateway
Protocol (BGP). The operators of each network need to have
control over the flow of traffic through the AS. However, BGP
does not facilitate common traffic engineering tasks, such as
balancing load across multiple links to a neighboring AS or
directing traffic to a different neighbor. Solving these
problems is difficult because the number of possible changes
to routing policies is too large to exhaustively test all
possibilities, some changes in routing policy can have an
unpredictable effect on the flow of traffic, and the BGP
decision process implemented by router vendors limits an
operator's control over path selection. In this talk, we
demonstrate that it is possible to predictably model the
changes in traffic flows in response to BGP policy changes,
given that policies are adapted in a certain fashion. Based
on analysis of routing tables and traffic measurements from the
AT&T backbone, we show that operators can control the scale
of the traffic engineering problem by focusing on the small
fraction of destination prefixes (and sets of related prefixes)
responsible for the majority of traffic. Furthermore, they can
make the effects of their changes more predictable by following
specific policy guidelines and selecting configuration options
that make the BGP decision process deterministic. This allows
an operator to gain more control over network traffic within
the existing BGP framework.
Copyright. All Rights Reserved.