Re: Oracle Replication Server

From: Scott Martin <smartin_at_bluewhale.emergent.com>
Date: 1995/05/20
Message-ID: <3pljak$rnp_at_kadath.zeitgeist.net>#1/1

In article <3ph8e0$3tf$1_at_mhadf.production.compuserve.com>, Robert Gauf <75301.2763_at_CompuServe.COM> wrote:
>The "Replication Server" is indeed separately licensed. Oracle
>saw fit to recover some of its research dollars (yet again ;-) )
>It consits of a bunch of PL/SQL procedures and some other extra
>stuff (like asynch procedure queues and a whack of triggers).
>We have done some small tests, somewhat successfully using the
>7.1.4 read-only snapshots. We are currently testing 7.1.6
>updatable snapshots before commiting live production databases
>to this technology.
>

We have developed a tool, a derivative of our redo log analyzer, SQL*Trax, which can do replication from a parent database to a child database by building a SQL script from the changes found in the Oracle redo logs. Note that, as the replication is done through SQL, the child database need not be Oracle.

This tool has many limitations over the solutions provided by Oracle but it has one major plus -

• NO PERFORMANCE DEGREDATION ON THE -------- ---- PRODUCTION DATABASE DURING BUSINESS HOURS ----

This is, of course, because the information needed for the replication is found in the redo logs which are already being maintained by Oracle. In many ways, this is a much better architecture than using triggers and PL/SQL. In fact, most other DB vendors who do replication use a log based approach.

We are currently looking for a Beta site for this product (preferrably in the San Francisco Bay Area). If anyone is interested in more information on this product, please do not hesitate to call me at 415-851-9576 or E-Mail to smartin_at_turnstone.com.

SCott.

>Your local rep should have a White Paper from Oracle on 7.1.6
>features and another on Database Replication that was passed out
>at IOUW '94.
>
>Hope this helps.
Received on Sat May 20 1995 - 00:00:00 CEST