Recovery in distributed extended long-lived transaction models

We address the recovery and the rollback problem in distributed collaborative transactions. We propose a solution to the problem in a generalized ARIES framework. We modified its existing data structures and provided additional data structures for recovery of distributed extended long-lived transactions. In the proposed model the transactions communicate and collaborate only by exchanging messages. The messages are logged along with usual database actions. In recovery of distributed extended transactions these message logs and message tables are extensively used. The recovery algorithms work in a distributed environment, under extended transaction models, with different kinds of failures and transaction rollback.