Technical Summary
Disruptions in end-to-end path connectivity, which last longer than
one retransmission timeout, cause suboptimal TCP performance.
This document proposes an algorithm to make TCP more robust to long
connectivity disruptions (TCP-LCD). It describes how standard ICMP
messages can be exploited during timeout-based loss recovery to
disambiguate true congestion loss from non-congestion loss caused by
connectivity disruptions.
Working Group Summary
Nothing exceptional occurred during the working group process for this
document.
Document Quality
There is an existing implementation in Linux, which was worked on by
the document authors. This has been in the vanilla kernel since version
2.6.32 (for about a year) and always enabled. Several Linux distributions
have shipped with this and no negative experiences have been reported that
were isolated to this mechanism.
Personnel
Wesley Eddy (Wesley.M.Eddy@nasa.gov) is the document shepherd. Lars
Eggert (lars.eggert@nokia.com) has reviewed this document for the IESG.