Blocking Reduction in Two-Phase Commit Protocol with Multiple Backup Sites

Abstract

The two-phase commit (2PC) protocol (or its variation) is widely employed for commit processing in distributed data base systems (DDBSs). The blocking phenomena in 2PC reduces the availability of the system as the blocked transactions keep all the resources until they receive the final command from the coordinator after its recovery. To remove the blocking problem in 2PC, three phase commit (3PC) protocol was proposed. Although 3PC protocol eliminates the blocking problem, it involves an extra round of message transmission, which degrades the performance in DDBS (Internet environments).

To reduce blocking, we propose a backup commit (BC) protocol by attaching multiple backup sites to the coordinator site. In this protocol, after receiving responses from all participants in the first phase, the coordinator communicates the final decision to the backup sites in the backup phase. Afterwards, it sends the final decision to the participants. When blocking occurs due to the failure of the coordinator site, the participant sites can terminate the transaction by consulting a backup site of the coordinator. In this way, the BC protocol achieves non-blocking property in most of the coordinator site failures.

The BC protocol suits best for World Wide Web (or Internet) environments where a server has to face high rush of electronic commerce transactions that involve multiple participants. Also in the Internet environment, sites fail frequently and messages take longer delivery time. In this situation with extra hardware, the BC protocol reduces the blocking problem without involving expensive communication cycle as compared to 3PC. Through simulation experiments it has been shown that the BC protocol exhibits superior throughput and response time performance over the 3PC protocol and performs closely with the 2PC protocol.