Ingres Corp, Emeryville, California has come out with Ingres/Replicator, a database server designed to enable organisations to distribute, copy and maintain data in multiple locations by replicating the information. The Ingres/Replicator incorporates a master-slave architecture. The master or central database server stores the original information and its duplicate is the slave. There can be more than one slave in the architecture, but the first slave is responsible for propagating the rest of the databases operating in a local area network. In the past, organisations have used two-phase commit database technology to distribute information. Two-phase commit is basically an all-or-nothing scenario, data has to be replicated and distributed in a synchronised format and on a time basis. This reduces information availability and can prove expensive, says Ingres’s UK product engineering manager, Ian Howells. For example, when updates occur to a table in a database, the entire table has to be replicated and distributed to different databases on the network. However in a replication architecture, users are able to replicate and deploy only the data that has been updated. Other Ingres/Replicator features include data integrity, which ensures that data replication only takes place after a transaction has committed locally – and in entirety; and fault-tolerance, which enables users to switch to any other database automatically if one database fails. The database is currently in beta test and will be available in 90 days, but no prices were given. Ingres/Replicator supports all ASK Group Inc development tools, as well as Ingres third party tools. It also supports non-Ingres databases including IBM Corp’s DB/2 and IMS, Digital Equipment Corp’s RMS and Rdb databases and Hewlett-Packard Co’s AllBase, the company notes.
