Constructing Transaction Serialization Order for Incremental Data Warehouse Refresh

In typical practice of data warehouse, the warehouse data is kept in a separate site from the operation data. Changes to the operation data to propagated to the warehouse site periodically, usually by shipping the new log entries in the operation site. When the operation site is a parallel a parallel DBMS, it is necessary to arrange the log entries into an coherent order for the warehouse data to maintain consistency and integrity. This paper solves the problem by presenting a simple method for constructing a global transaction order out of a set of local logs of a parallel or distributed system. For serial database systems, transaction order can easily be inferred from local database logs. For parallel or distributed database systems, since each local log les has only partial information of each transaction, this problem becomes diiculty. In this paper, we present a simple method that determines the order of transactions by examining only the relative position of their vote and commit entries, and needs not looking into what data items each transaction accesses. Based on this mechanism, we present a method that eeciently constructs a global transaction order. Conceptually, our method rst connects the vote and commit symbols of local log les into a network. Then it traverse the network to detect inconsistency in the ordering of commit symbols in diierent log. It then resolves the inconsistency by adjusting the position of commit symbols in the network. Our method has the following features. First, it examines only the vote and commit entries of transactions, but not the data accessing entries. Second, for vote or commit 1 entries, only their relative positions and their transaction ID (to identify from which transaction an entry comes) requires attention. No other information recorded in the vote or commit entry needs be examined. Finally, it has low complexity. For a set of local logs with l totally, only O(l) cost is required both in terms of time and space.