Comments on: Disruption versus chasm crossing in the database market http://www.dbms2.com/2008/03/28/disruption-versus-chasm-crossing-in-the-database-market/ Choices in data management and analysis Fri, 29 Aug 2008 03:42:08 +0000 http://wordpress.org/?v=2.5.1 By: Daniel Weinreb http://www.dbms2.com/2008/03/28/disruption-versus-chasm-crossing-in-the-database-market/#comment-80124 Daniel Weinreb Sun, 30 Mar 2008 16:02:59 +0000 http://www.dbms2.com/2008/03/28/disruption-versus-chasm-crossing-in-the-database-market/#comment-80124 I think the report is right when it says that there are a lot of conservative database users on whom it will take more time for open-source DBMS's to make a good impression. I agree that users are concerned about service and support; but many people I've talked to are pretty unimpressed with Oracle's support, and we'll see how good EnterpriseDB's support is, over time. About performance, many, many applications don't need the kind of performance that mainstream DBMS providers worry about, such as winning the official TPC benchmark wars. In "The Innovator's Dilemma", the classic response of the dominant old-line guys to the new upstart guys is something like: What, are they kidding? That thing can't succeed at what our customers need! It's too small, too slow, doesn't have all the features. We know from experience that that's what customers need. We can safely ignore these guys. But the old-line guys are sometimes blinkered in their view of the greater market, and the segments that are different from their own favorite target customers. In this case, I think there are a lot of applications for with good open source DBMS's are well-suited. I saw this same thing happen when I was at BEA. At first, JBoss was treated as a joke. Standard J2EE app server stress-test benchmarks could not even be done, because JBoss would crash! This led to much mirth at BEA. But JBoss continued to improve, at impressive speed. Many people could use a J2EE app server who had more modest needs and didn't want to pay much. BEA was not addressing this market nor thinking about its needs very much. And then, BEA's big customers started to say that they'd be using JBoss for at least some of their future application development! Then BEA started to take notice. One reason the customers gave was that they liked to be able to read the source, both to understand better what was going on inside the app server, and to diagnose bugs or problems themselves. BEA considered allowing some customers to have source access; they considered going into the JBoss area themselves; they considered interoperating; and while I was there, they did so much considering for so long that they lost valuable time. I don't know if they've settled on a strategy since I left. (There are other open-source J2EE app servers now, too.) I agree that they're not dead ducks. They get most of their revenue from license upgrades and service, and although those may fall off, it would take a long time. I think the report is right when it says that there are a lot of conservative database users on whom it will take more time for open-source DBMS’s to make a good impression. I agree that users are concerned about service and support; but many people I’ve talked to are pretty unimpressed with Oracle’s support, and we’ll see how good EnterpriseDB’s support is, over time.

About performance, many, many applications don’t need the kind of performance that mainstream DBMS providers worry about, such as winning the official TPC benchmark wars.

In “The Innovator’s Dilemma”, the classic response of the dominant old-line guys to the new upstart guys is something like: What, are they kidding? That thing can’t succeed at what our customers need! It’s too small, too slow, doesn’t have all the features. We know from experience that that’s what customers need. We can safely ignore these guys.

But the old-line guys are sometimes blinkered in their view of the greater market, and the segments that are different from their own favorite target customers. In this case, I think there are a lot of applications for with good open source DBMS’s are well-suited.

I saw this same thing happen when I was at BEA. At first, JBoss was treated as a joke. Standard J2EE app server stress-test benchmarks could not even be done, because JBoss would crash! This led to much mirth at BEA. But JBoss continued to improve, at impressive speed. Many people could use a J2EE app server who had more modest needs and didn’t want to pay much. BEA was not addressing this market nor thinking about its needs very much.

And then, BEA’s big customers started to say that they’d be using JBoss for at least some of their future application development! Then BEA started to take notice. One reason the customers gave was that they liked to be able to read the source, both to understand better what was going on inside the app server, and to diagnose bugs or problems themselves. BEA considered allowing some customers to have source access; they considered going into the JBoss area themselves; they considered interoperating; and while I was there, they did so much considering for so long that they lost valuable time. I don’t know if they’ve settled on a strategy since I left. (There are other open-source J2EE app servers now, too.)

I agree that they’re not dead ducks. They get most of their revenue from license upgrades and service, and although those may fall off, it would take a long time.

]]>