I want to share my excitement with you about an event that just took place in the Postgres community.
A couple of days ago, the Postgres enhancement known as HOT (Heap Only Tuples) was committed into the code base for the next release (v8.3). In a nutshell, HOT significantly increases the performance of Postgres over long periods of time in update-intensive OLTP applications.
“This is a good thing,” you might say, “but what's the big deal?” I’m glad you asked...
HOT is the largest and deepest single change to the Postgres base that has been implemented in the last several years. It impacts many areas of the core database, and is truly something that could be called "rocket science." The reason I'm so pleased about HOT (besides the obvious customer benefit) is that the enhancement was conceived, designed, implemented, and shepherded through the community entirely by the EnterpriseDB team.
The acceptance of HOT into the 8.3 code base means two things: First, it is the flagship example of our commitment to contribute back to the community large quantities of enormous value. It feels great to give back to a community that has given us the basis of our very existence. Second, HOT firmly seats EnterpriseDB as a leader in the community; one of the very few organizations — perhaps the only one — on the planet that is able to deliver Postgres technology of this complexity and size.
It is a particularly great day for the team that did the work. First and foremost, I have to recognize Pavan Deolasee, who was the thought leader on this effort from beginning to end. Of course, many others contributed greatly to the project, including Anupama Aherrao, Korry Douglas, Dharmendra Goyal, Jonah Harris, Hope Jiang, Sivakumar Krishnamurthy, Heikki Linnakangas, Shoaib Mir, Bruce Momjian, Nikhil Sontakke, Greg Stark, Yaser Raja, Umair Shahid, and Peter Yarrow.
Thursday, September 27, 2007
Tuesday, September 25, 2007
Home Sweet Home
Over the Labor Day holiday weekend, EnterpriseDB (finally) moved into its new offices. In 2 years, we’ve gone from 1,000 feet in the the back of a warehouse, to a 5,000-square foot sublet, to an 8,000 square foot sublet, and now finally to a 17,000 foot, 5+ year lease that was built to our specifications. We now have plenty of space for guests, conferences, training, and more. I’ve included some pictures below.
So if you’re a friend of EnterpriseDB, and you’re in the Edison, New Jersey area (one block from the Metropark Amtrak station), feel free to drop by for a cup of coffee and a dime tour. And if you need an office to use for the day, just ask.
So if you’re a friend of EnterpriseDB, and you’re in the Edison, New Jersey area (one block from the Metropark Amtrak station), feel free to drop by for a cup of coffee and a dime tour. And if you need an office to use for the day, just ask.
Thursday, September 6, 2007
A New Database Blog
Michael Stonebraker and several other database luminaries have just started a new blog that is worth checking out. Stonebraker was the principal architect behind Postgres and Ingres, and commercialized them into Ingres Corp., Illustra, and others.
The first entry in the blog calls attention to Stonebraker's new company Vertica and its column-oriented database technology. For data warehousing and certain specialized databases, Stonebraker claims a 50-fold performance improvement. Very impressive, indeed. And for these specialized circumstances, probably very appropriate.
Stonebraker makes the point that traditional databases, including Oracle and Postgres, are 25 years old and somewhat long-in-the-tooth. Perhaps. But for most of us, they are also proven and stable technologies that work for the millions of systems in which they're installed. Indeed, it is precisely because Oracle and Postgres are so proven and stable that EnterpriseDB believes they should be far less expensive and cross-compatible with each other.
Vertica's job is to take databases to the next frontier. Ours is to make the current frontier easy to use, affordable, and accessible to everyone.
The first entry in the blog calls attention to Stonebraker's new company Vertica and its column-oriented database technology. For data warehousing and certain specialized databases, Stonebraker claims a 50-fold performance improvement. Very impressive, indeed. And for these specialized circumstances, probably very appropriate.
Stonebraker makes the point that traditional databases, including Oracle and Postgres, are 25 years old and somewhat long-in-the-tooth. Perhaps. But for most of us, they are also proven and stable technologies that work for the millions of systems in which they're installed. Indeed, it is precisely because Oracle and Postgres are so proven and stable that EnterpriseDB believes they should be far less expensive and cross-compatible with each other.
Vertica's job is to take databases to the next frontier. Ours is to make the current frontier easy to use, affordable, and accessible to everyone.
Subscribe to:
Posts (Atom)