Murray’s Eight Phases to a Totally Distributed System
Phase 1: The first phase is characterized by host-based, real-time query and update. This phase is traditional on-line information system processing, where dumb terminals access host-based applications to view and update data
Phase 2: The second phase provides additional query capabilities through file transfers to PCs.
Phase 3: The third phase adds batch updating form PC data. This phase reverses the philosophy of Phase 2 by making the PC database the master.
Phase 4: The forth phase enables real-time query and update from either host or PC. This phase extends the capabilities of the PCs by allowing them to update the host on-line.
Phase 5: The fifth phase introduces homogeneous cooperative processing without two-phase commit, that is, like databases run on the same hardware and system software platforms. This phase adds true distributed databases, across similar or identical platforms.
Phase 6: The sixth phase moves to heterogeneous cooperative processing without two-phase commit, that is, databases run on a mix of platforms. This phase extends the previous one by permitting distributed databases across mixed platforms.
Phase 7: This seventh phase adds the all-important two-phase commit capability (to homogeneous databases), going a system a true distributed database.
Phase 8: This phase extends Phase 7 to heterogeneous databases.