Xref: alice comp.databases.oracle.server:56345
Path: alice!news-feed.fnsi.net!netnews.com!news-peer-europe.sprintlink.net!news.sprintlink.net!news.algonet.se!algonet!newsfeed.online.be!uunet!ams.uu.net!ffx.uu.net!ffx4nh4!not-for-mail
From: robertt@ccslink.com (Robert Toellner)
Newsgroups: comp.databases.oracle.server
Subject: Replication w/ multi dissimilar db's
Date: 7 Jul 1999 14:22:08 GMT
Organization: CCS
Lines: 34
Message-ID: <8DFC5F0FBroberttccslinkcom@news1.server.wcom.com>
X-Trace: ffx2nh4.news.uu.net 931357328 4283 207.247.70.66 (7 Jul 1999 14:22:08 GMT)
X-Complaints-To: news@ffx2nh4.news.uu.net
NNTP-Posting-Date: 7 Jul 1999 14:22:08 GMT
User-Agent: Xnews/2.07.01

I hope someone can help - this may be easy but I do not know Oracle's 
capabilities.

I need to set up a read only ( for query purposes ) database and I'm 
considering an Oracle 8i server. This database will simply be a datastore ( 
datawarehouse? ) of vasts amount of data to be displayed to 
internet/intranet users.

The data, however, needs to be obtained and regularly synchronized ( 
nightly ) from multiple dissimilar databases. We're talking probably 20 to 
100 different db's all of which may be of any manufacturer/type (ie dBASE, 
Oracle, Sybase, Informix, text delimited, etc ) and residing on possibly 
different server platforms (NT, Win95, AS/400, Unix/AIX, etc).

Does Oracle supply connection tools that would make the replication with 
all these DB's? Or, any other manufacturer?

I would also like the web db to control the replication of data if 
possible. Naturally, I would rather setup the connection parameters at the 
one main server instead of setting the connection at all the underlaying db 
servers.

Also, if someone could lead me on how Oracle actually performs the 
replication. The initial implementation of the web db will be huge - but 
after that, the replication process should only pick up 0 - 10000 row 
changes per underlaying db server. For replication, does something? need to 
reside on both the web db and the to-be replicated db's??

My goal is to have the web db remain independent and perform all the 
necessary logic needed to get a current set of data from all the other db 
servers with as little as possible ( none actually ) impact on those db 
servers.

Thanks