Post of the Day
March 30, 1998
From our AOL
Subject: Re: Happy, Fat and Content.
Sorry for the icepick. It was a crude attempt at humor and I've gotta learn to stop that. It's just so rare that anyone arrives at a party and hands out a pop quiz, while wearing a look that says I know you're all gonna flunk this. I was mistaken and regret the offense.
"So, what I really want to know is what is this companies vision for databases and software applications in a networked society. Would appreciate any feedback."
I think the vision goes pretty much like this:
In a networked household [read: organization, society, galaxy, whatever] there is a very compelling technical argument for centralizing certain kinds of computer processes. Storing and fetching shared data is one of these. Just as it made more sense at the dawn of the database era to separate application logic from data management functions and let a specialist program (the "database") handle the data, it has become apparent in the networked era that network administration gets a whole lot easier if you can concentrate these database activities into fewer places than there are network participants. (To use Ellison's illustration, imagine the mess that would result if telephone companies stored a copy of their electronic phone directory data at each network address where there's a telephone. For any new phone number added to the directory, millions and millions of updates across the network would be necessary. Eventually, directory maintenance traffic would grow to such a volume that it started crowding out regular phone calls.) That's because, among all the things we do with computers -- networked or otherwise -- providing shared simultaneous access to information while it's being updated and administered is a major engineering feat, one that involves complex record locking and transaction control algorthms and incredible efficiency reading and writing indexes. Doing this with an already-active communications network standing between a user and his data just makes it an order of magnitude more complicated.
So, the answer to your question is this: far more than any company in the worldwide software industry, Oracle understands the complications that arise when you place one of these complex database applications into network service and "scale it up" (i.e., begin to add users and applications competing for attention). The company intends to remain the worldwide leader in this specialized technology and to leverage it as the foundation of powerful new products such as the Oracle Applications, Media Servers, and so on. Contrary to wacky reports that the need for database software is already satieated, Oracle believes that the market is still in its infancy. Like nearly everyone else who has been touched by the evolution to network computing, the company anticipates a time when billions of people will benefit from online database access and it intends to retain a pretty substantial grasp on the greatly enlarged market that will represent.
As for the NC hardware, I think Powerski8 summarized that very well. As a leader in the discussion of computing's next wave, Oracle needed to demonstrate the feasibility of low-cost network appliances and, never shy, Larry Ellison "put his money where his mouth was". As a result, IBM, Microsoft, and Intel quickly announced network computing strategies of their own, which has to leave Larry feeling pretty giddy if we assume that his real motive was to legitimize an alternative to the Wintel PC and to upset the economics beneath the old PC architecture. Whether Oracle's Network Computers division ever becomes a going concern or not, it's starting to look like a brilliant move: REDHERRING magazine recently predicted that network computers and settop boxes will surpass PC shipments by 2005.
And if LJE has his way, every one of those non-Windows "computers" will need to access data somewhere. At least a little of the giddiness you've observed here includes a few of us who think he'll be ready.