Re: [pgsql-advocacy] Oracle buys Innobase

Chris Travers <chris@travelamericas.com>

From: Chris Travers <chris@travelamericas.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: "Jim C. Nasby" <jnasby@pervasive.com>, "Joshua D. Drake" <jd@commandprompt.com>, Jeffrey Melloy <jmelloy@visualdistortion.org>, Bruce Momjian <pgman@candle.pha.pa.us>, Matthew Terenzio <matt@jobsforge.com>, PgSQL General <pgsql-general@postgresql.org>
Date: 2005-10-17T21:43:27Z
Lists: pgsql-general
Tom Lane wrote:

>
>
>They might hope that they could drive the existing support companies out
>of business (assuming they didn't get convicted of antitrust violations
>first --- which would be an open-and-shut case, but with the Republicans
>in office they probably wouldn't get prosecuted :-().
>
Sort of off-topic but after actually reading up on notable antitrust 
cases (such as AT&T), I think that the current situation wrt Microsoft 
is exactly how antitrust law works best in our legal system.  And anyone 
who thinks that Microsoft effectively put this behind them has not been 
following Novell v. Microsoft and a myriad of other cases.  Because 
Microsoft has lost their case, certain facts cannot be litigated, 
meaning that it is now open season on suing Microsoft for antitrust 
violations.... (IANAL, but you can ask one about "collateral estoppel" 
which is a really nasty ball of wax for Microsoft at the moment).  The 
current settlement really was forced on the DoJ by the appeals court.  
This is one area where Microsoft would have been better off (and we 
might be worse off) had they been broken up.  For example, AMD faces a 
much harder antitrust case against Intel than Novel does against 
Microsoft for the reason that Intel has settled all previous antitrust 
cases without admitting guilt.  Don't think so?  Why do you think 
Microsoft settled IBM's antitrust claims before the lawsuit was even 
filed (normally people at least go through pretrial motions to see how 
much of the complaint they can get dropped before settling)?  Indeed I 
have personally wondered if Microsoft opened themselves up to more 
lawsuits by recommending that Baystar invest in SCO....

As for Oracle, they don't exactly have a steller reputation.  However, 
they hopefully have enough to sense to avoid antitrust cases they could 
lose.  At least in the past, their prior questionable actions have been 
of generally unfair business practices such as industrial espionage 
(didn't they hire the firm that got the janatorial contract at Microsoft 
to do dumpster-diving for them in 1999 or so).

But the biggest issue for them is that other parties (such as IBM and 
Microsoft) have been making substantial inroads into Oracle's core 
market.  Spend too much in the way of resources attacking us and they 
divert resources from the clear dangers that they have from large 
commercial competitors.

>  Then they raise
>their rates to make lotsa money, or maybe they'd think they could drop
>support at that point and the project would die for lack of commercial
>support.  (They seem to understand open-source poorly enough that they
>might think that would happen.)
>  
>
Who knows?  Maybe they will resort to dumpster-diving to try to discover 
our super-secret-source-code... ;-)

>I don't see any of this happening though.  As suggested upthread,
>the very *last* thing Oracle wants is to raise the visibility and
>credibility of Postgres by a couple of orders of magnitude --- which
>is exactly what they'd be doing by offering support for it, even if
>the support was only temporary.  The effects of getting the word out
>would persist long afterwards.
>  
>
I would suggest that Oracle has not formed a strategy for outcompeting 
us yet, and it may be several years before they take the threat we pose 
seriously enough to really start work on it.  I would suspect that we 
are treated as "one of a crowd of mid-size RDBMS competitors," and have 
not been singled out yet for special treatment (MySQL was singled out in 
2000 at the latest).  Oracle's current strategy seems to be in trying to 
push things like parallel queries, grid computing, etc. as a way of 
providing scalability and room for growth, providing advantages for 
certain scenarios.  They could then use the high-end to subsidize the 
low end, like Sun does with Solaris (though I think that this is a 
losing strategy and Microsoft's inverse strategy of subsidizing the 
high-end with the commodity market is ultimately more effective).

Maybe when Bizgres MPP comes out things will change ;-)

Best Wishes,
Chris Travers