The First Wives Club
July 31st, 2008
No doubt you saw the press coverage last week of Microsoft’s acquisition of DATAllegro. I’ve received dozens of emails asking what I think of the acquisition, particularly since the DATAllegro executives have been stating in the press and in their blogs how SQL Server is better than Ingres.
What do I think? I feel like Ingres is the wife who stood behind her husband as he built his successful career, who got publicly dumped for a trophy wife. Move over Ivana Trump and make some room for Ingres in the First Wives Club!
It’ll be interesting to see DATAllegro go through a transformation over the next 18 months but instead of the traditional makeover of hair plugs, a new wardrobe, an eye lift and a fast car, we expect to see it emerge with a new database, operating system and hardware platform. I jest of course! I have no doubt that the relationship between Ingres Corporation and DATAllegro will remain cordial as there are some very important customers running the DATAllegro/Ingres solution to consider.
Why didn’t DATAllegro build their solution on a Microsoft stack of Windows and SQL Server to begin with? Well, you’ll have to ask Stuart and Mark that question, but I suspect that for a startup, cost would have been one factor. But more importantly, without access to the source code for the database and operating system it would have been impossible for DATAllegro to build a solution that could scale to meet the requirements of the most demanding data warehouses.
I find it amusing that the proprietary vendors keep having to “buy innovation” from the open source side despite the hundreds of millions they pour into research in their own labs each year.
July 31st, 2008 at 11:12 am
This acquisition does not make sense to me. If Microsoft plans on replacing Ingres with SQL, and some Linux distribution with Windows - why not just develop their own warehouse product in the first place?
July 31st, 2008 at 11:28 am
In the words of Ivana Trump “Don’t get even, get everything!”
July 31st, 2008 at 3:18 pm
“why not just develop their own warehouse product in the first place?”
Because it’s not that simple. A lot of the difficulty is in the details; learning how to get a user query decomposed into a series of distributed queries, and get it right. When do you redistribute? when do you repartition? do you stream or do you build a temp result set and move it? let’s throw a full join into the mix, do you get the right answer? What’s the right cpu, disk, interconnect balance? and on and on.
The Datallegro engine guys spent a few years getting all of that stuff figured out. That’s what MS acquired.
Emma, I like the first-wife analogy.
I will note that the amount of PR-mandatory Ingres-bashing in the acquisition PR was about as low key as you can get in the circumstances. That ought to tell everyone something as well — this deal wasn’t about Ingres’s weaknesses, it was about adding a high end to SQL Server. Roy Hann made a good point in his comp.databases.ingres post: the next venture to take a shot at an open-source based data warehouse can make a safe bet by choosing Ingres for the DBMS.
August 1st, 2008 at 5:26 am
Classic!