November 15, 2006
Objects and Databases: State of the Union 2006William Cook, Moderator: Let's change the focus a little bit. We had dinner last night and I asked some gentle questions to get to know the people a little better. One of the things that kept coming up is that the issues we're really ought to be talking about are not technical issues perhaps entirely, but our business and social issues, so we have a couple of questions along those lines.
"How can you pursue object orientated databases when corporate customers insist their data be in a relational database?" and "We have these databases -- what are the problems we see inthe industry?" Because what is really happening is all this is being played out on the desktops of DBAs and programmers in the industry.
What's the social and business climate, how is that affecting you guys?
Derek Henninger, Progress Software: I want to relate to that, a comment earlier was that we as programmers need to fight back. I think our chance to do that was the dotcom boom, at least what we saw that was when DBAs had less power and anything to get that new app on the Internet was acceptable. What I see is the pendulum has swung back to the DBA and the core CIO and all that infrastructure they built.
I believe that the relational world is going to continue for a while, I also believe strongly in the value and benefits of object databases therefore I see no alternative but improve mapping and the transformation tools around that need to become a lot more sophisticated. It's not just simply OR mapping anymore. I think its object to object mapping as well. As you look at twodifferent applications maybe written with a Versant database and an ObjectStore database and a Gemstone database or a db4o database. At some point we're going to want to map between them. So I do see that's going to be an important technology advance in order to make all this practicable, not to mention relational and hierarchical and all these other databases.
Patrick Linskey, BEA: I think I agree a lot. One of the things that OR mapping does a really good job of, is organizationally, lets people get along in the organization rather than kind of Negotiation 101 is that when you sit down on the negotiation table, you're not trying to win; you're trying to come to a solution that everyone can agree to. That's where I think OR mapping does a pretty good job of, in a lot of companies where you have DBAs say "Thou shall use that schema", if you want to add something put it into the top of the pipe over here and then two years later we will tell you whether or not we will accept your change.
What I've seen OR mapping do over the years is, companies say: "We can do that it's just"going to cost us this much to set up this mapping" or "It's going to be a little bit slower" or "We need to write a little bit of custom Java code" or whatever the case may be. It puts the developers and the DBAs in a position where instead of using political muscle to create a winner in a given situation you actually get to blame a vendor and work together with your DBAs and developers and say "What is it what we want to be able to do? What's the datamodel, what's the object model, how do we make them bridge?". Then you go around and you see the different people who are pitching you on software. How can you make this happen? I think that's essentially why OR mapping is popular, because there are so many people who must use their schema that is used by 70 applications over the course of the last 15 years. They can't change it, so but they want it to behave differently, they don't want to fight that battle anymore.
Robert Greene, Versant: How many of you have seen the size of an object relational mapping layer? It's pretty enormous, and that has to execute on a CPU and that's a lot of instructions and there's just no getting around that, the cost involved in using that approach.
One of the things that is going to be championed up into the upper management in some of these companies is the cost of computing. If you look at the message that is coming out of companies like Sun, when they start talking about the fact that it's going to cost you more to run your computer for the life of that computer than it is to buy it, then you need to startlooking at things like when I choose to use a relational database for this application, if it means that I need to run on twice or three times as many CPUs as I would be running if I chose a different technology, is that the right decision we're making?
If you look at some of these applications out there, I just happen to be working a lot in the massive multiplayer online gaming space over friends back to Second Life again, but they're talking about 10 million users and 150,000 servers by 2008.
So imagine if choosing one technology over another at the database layer meant that you could use 70,000 servers instead of the 150,000 servers. That's the kind of thing I think that IT management....if they're good at what they do...ought to be looking at when making those decision instead of just "No my guys know Oracle and that's it". So we're an Oracle shop we're going to use it blindly without a proper analysis.
Bob Walker, Gemstone Systems: I think that's really true. And one of things you should keep an eye on is "Where's the money? Where's it going?" There's a faction in industry that is well vested in selling more hardware, networks -- at odds with optimizing the software. 
Previous Page |
1 Introduction
|
2 Bob Walker, Gemstone Systems
|
3 Derek Henninger, Progress Software
|
4 Robert Greene, Versant
|
5 Erik Meijer, Microsoft
|
6 Christof Wittig, db4objects
|
7 Patrick Linskey, BEA
|
8 Craig Russell, Sun Microsystems
|
9 Ten Years from Now?
|
10 Data -- It's Scary
|
11 Acronym Soup
|
12 Business and Social Issues
|
13 Back to the Social Question
|
14 Different Viewpoints
|
15 A Bigger Picture Question?
|
16 A Service-Oriented Impedance Mismatch?
|
17 Wrap Up
Next Page
|
 |
|
|||||||||||||||||||||||||||||
|
|
|
