Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Sybase <-> Oracle Replication

Sybase <-> Oracle Replication

From: Alex <avilner_at_gmail.com>
Date: 6 Apr 2005 20:49:51 -0700
Message-ID: <1112845791.327881.246150@g14g2000cwa.googlegroups.com>


Dear netters,

We are in the process of migrating some of our applications from Sybase (12.5) to Oracle (10g), and are looking for a transitional plan for keeping the systems running in parallel.

Specifically, we plan on having 3 database servers (simplified scenario) as follows:

Server #1 (Sybase) - main databases
Server #2 (Sybase) - supplemental databases
Server #3 (Oracle) - migrated main databases (mostly, replica of #1)

Sybase is runing on Solaris, Oracle will be running on AIX.

In our current environment, we need to maintain the following data movements:
Server #1 -> Server #3 (one way).
This one is for when both servers go live and we begin migrating applications to be consumers of data in Oracle, rather than Sybase. Existing applications will continue working with Sybase until they are migrated, and data is also populated on the Sybase server through a number of data feeds. For the time being, Oracle Server #3 will be a read-only database, that needs to be updated on a regular basis, in near real-time (up to 1 hour delay is acceptable). Anticipated volume is several million rows of data per day, more during peak times, but otherwise -- spread throughout the day. Data structures between Sybase #1 and Oracle #3 will be fairly similar after the migration.

Server #1 <-> Server #2 (two way).
In addition to getting external feeds, server #1 will also be receiving data from Server #2. The volume of data is not large -- data on Server #2 is modified either through some batch processing (up to 100,000 rows per day) or via manual data entry (mostly, corrections, etc., up to about 10,000 rows per day). The data movement also allows for some latency, up to 1 hour is acceptable.
Data structures are different between Server #1 and Server #2, some massaging/transformation is required.

We are looking to implement some sort of replication to fit both of these requirements. For options we are looking at:
- Sybase Replication Server + DirectConnect to Oracle

In either case, replication will be for changes made since last replication was run. There may be additional conditions imposed on what gets replicated, further limiting the amount of data to be replicated. A complete refresh of data is not an option, due to large overall volume of data stored in all databases.

What are the recommendations? Do you know if anybody has evaluated these (and other) options for pros/cons -- as far as ease of implementation, cost, database changes' requirements, additional hardware requirements, etc.?

Thank you in advance!

--Alex Received on Wed Apr 06 2005 - 22:49:51 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US