Masking Failures of TCP Servers
Without Changing almost
Anything
Keith Marzullo joint work with Dmitrii Zagorodnov, Thomas Bressoud (Lucent Bell
Labs) and Lorenzo Alvisi and Ayman El-Khahsab (UT Austin)
I will describe the design and implementation of a fault-tolerant TCP
(FT-TCP) that allows a crashed server to keep its TCP connections while
it either recovers or it is failed over to a backup. The failure and recovery
of the server process are completely transparent to client
processes connected with it via TCP. FT-TCP does not affect the software
running on a client, does not require to change the server's
TCP implementation,
and does not use a proxy. The overhead of using FT-TCP is very low in a
reasonable network configuration.