"New Breed" Database Extensibility

At present, I can think of at least 5 "new breed" database vendors that you allow you to extend their SQL language in some form or another:

Netezza
ParAccel
Kognitio
Greenplum (I think)
DATAllegro/Microsoft

Of course, the old guard is well-represented in this category as well - Oracle, DB2, SQL Server, Teradata, etc. all allow language extension via custom functions, plug-ins, etc.

I don't know enough about it yet to know, but I think that Greenplum and AsterData might also belong on this list due to their support for Map/Reduce.

It wasn't all that long ago (3 or 4 years) that most of these vendors didn't even support the full SQL standard, never mind compiled-code extensions to their SQL language. Oh, the possibilities.

Netezza Acquires IISi

From the press release:

This acquisition allows us to seed the market for these third-party applications and will be a catalyst for creating a marketplace of offerings from our NDN members," said Jim Baum, president and COO, Netezza.

Now, everybody, repeat after me: Netezza is not a database company, Netezza is not a database company, Netezza is not a database company...

Playing Devil's Advocate: DATAllegro vs. Netezza

This recent article by DATAllegro CEO Stuart Frost is well worth reading (and not just because they quote me in it). If only because I love a good debate, I think there are a few counter-points worth making though.

Non-Proprietary Hardware


I love this section, particularly the quote from Intelligent Enterprise. Saying that DATAllegro has an "open, hardware-independent approach" reminds me of Henry Ford's opinion on available colors for the Model T: "You can have it any color you want as long as it's black". Yes, DATAllegro uses commodity hardware. Don't mistake that to mean that you have any say in what hardware will go into your DATAllegro system though.

Non-Proprietary Software


This section cracks me up too. If DATAllegro's software isn't proprietary, then what are their customers paying for?

Scalability


This section leaves me scratching my head a bit. Ok, DATAllegro systems scale to 400TB. Do they function well at that size, and what are they used for? For all we know, Netezza may be doing ad-hoc analysis at 200TB, while DATAllegro may be doing simple archiving at 400TB. I'm not sure the raw data size is ultimately what's important here.

Distributed DW


This section should get top billing. No, not because they quote me in it, but because this is where DATAllegro is set to clean house. Ok, so I guess that's an endorsement, not a counterpoint, but this concept will be huge, mark my words.

Parting Thoughts


Nit-picking aside, I think that it's really great that Mr. Frost is writing posts like this. They're admittedly biased, and a bit cheeky at times, but it's good to get some informed, inside perspective. I think DATAllegro would be better served to focus on their obvious strengths than to thumb-wrestle competitors on line items, but the verbal sparring is certainly more entertaining. As such, I eagerly look forward to more.

Netezza Acquires Predictive Analytics Company

Ok, so I know I'm a week behind the news here, but: last week, Netezza acquired NuTech Solutions, a "provider of advanced predictive analytics and optimization solutions" (official press release). I've been saying for a while that Netezza is no longer a database company, but I didn't think they'd prove me right this soon, or this way.

Thanks guys. :-)

My Definition of Appliance

There's been an increasing amount of discussion lately about what is and isn't an appliance, what types of appliances there are, etc. especially given today's announcement by Teradata. Personally, I don't quite understand what all the fuss is about, because it seems pretty simple to me.

In short, if I could assemble it myself - even if only in theory - it ain't an appliance.

This leads to some pretty easy distinctions:

Appliances
Netezza

Non-Appliances
Everything else

Just because it's pre-installed, pre-configured, low-maintenance and low-power doesn't make it an appliance. An appliance is all of those things, some by design and some by nature, but those are not the defining traits of an appliance.

So what is, and why does it mean you can't put it together yourself?

In a word: minimalism.

My dishwasher has in it exactly the parts it needs to be a dishwasher, with the least amount of connections necessary to make those parts work together for the intended purpose. The same is true for my washer, my dryer, my microwave and my toaster. Some of them share similar parts, but none of them have anything extra*, and none of them can be re-used as anything else. In short, to borrow words from the great Alton Brown, they are uni-taskers.

* Don't get all cynical on me and start making comments refrigerators with ice makers. Bells and whistles are a different animal.

While it makes imminent sense for database system vendors to use commodity hardware as much as possible, and even more sense for their customers to buy systems that use commodity hardware, the use of that commodity hardware means that their systems are not, by definition, as minimal as possible. And that, to me, means they're not appliances. Advanced, powerful, flexible, mind-boggling, adrenaline-generating and landscape-changing, yes. Appliances, no.

I can take a diesel engine, a 55-gallon drum and a space heater and make a pretty kick-ass dryer. It would be cheap (think bio-diesel) and the parts would be re-usable when I decided I wanted something different. Nobody I know would call it an appliance though.