(no subject)

In this paper, authors analyzed how topology and policy combine to
contribute to path inflation at three levels in the Internet: intra-
domain, peering, and inter-domain.

This paper uses trace-driven methodology to analyze the issue. Paper
started by measuring the POP-level backbone topology of 65 diverse
ISPs and their interconnections, by tracing from 42 vantage points to
all globally-routed IP address prefixes. This is challenging because
most ISPs are unwilling to share accurate, up-to-date information
about their topology or routing policy. However, authors devised
novel techniques to infer the intradomain routing policy and the
peering policy among pairs of ISPs.

At first, authors analyzed the Intra-domain impact on path inflation.
The research results showed that setting weights on intra-domain
links does not significantly impact path inflation, suggesting that
intra-domian traffic engineering is latency-sensitive. For the impact
of peering factors, authors firstly investigated the impact of
peering topology, the research suggested that the peering topology
does not inflate paths significantly. However, for impact of peering
policies, the results showed that peering policies often cause path
inflation, and these policies sometimes lead to highly inflated
paths. This means that they consume more network resources, probably
for both ISPs, than an optimal exit policy. At the third level, in
summary, the authors found that inter-domain routing in the current
Internet has a significant impact on path inflation with more than
half of the paths being longer than the shortest path. However,
policies such as no-vally and prefer-customer arising out of
commercial concerns are not a contributing factor. Instead, most of
the inflation comes from using shortest AS-path as the distance metric.

This paper is well written and presented with originality. Over the
past few years, researchers have found significant path inflation in
the Internet. This paper tried to answer a simple question: why are
Internet paths sometimes absurdly long? To investigate this issue, we
can get a broader understanding the factors that shape Internet
routes. This paper also is based on trace-driven analysis which could
give us a real figure of Internet. According to the result, a main
culprit of path inflation is the design of BGP, the current inter-
domain routing protocol, that makes it difficult to implement robust
topology-sensitive routing decisions. Thus, a question emerges that
need we redesign the BGP? How to tradeoff between the cost of
redesigning BGP and the impact of path inflation?
Received on Wed Nov 15 2006 - 16:45:26 EST