The Sun Also Rises

This cliché headline is one that Sun Microsystems has been waiting to read for many, many quarters. Specifically how many quarters, I don't know, but 4 or so years worth I think. When you were once a technology leader and suddenly you're losing market share and only ever turn in a profitable quarter by accident, you must begin to doubt yourself. I know from talking to Sun execs (some of them now-departed) that such clouds hovered over Sun's HQ in Santa Clara for quite a while. So, good luck to Sun CEO Jonathan Schwartz, who recently climbed into the hot seat and has been almost instantly rewarded with a significant gain in server market share (15.5 percent year-on-year growth).

It is a little early to conclude that Sun's problems are over, but the hurricane has passed and the buildings are still standing. Surprisingly, it now seems that the sick man of the server market, whom Sun has now overtaken, is Dell!

What happened?

Dell always was, and still is; Johnny-one-note. It's manufacturing culture drives the business and it's as Japanese as sushi served cold in Kyoto. Dell has spent years, in fact more than a decade, squeezing cost out of its manufacturing operation and honing its direct sales model. Over the years it has shaved billions from its costs and brought consternation to the competition. But things have changed.

Point one: Servers are no longer the commodities they once were and Dell's one note continues to sing “commodity product, commodity product, commodity product”.

Point two: In time, the law of diminishing returns ruthlessly diminishes the returns of any Johnny-one-note strategy. Dell cut its cost to the bone, so now all it has is bone. Some of its erstwhile customers are looking for service, but service means added cost, which is a little counter-cultural for Dell. Dell's marketing culture, it seems, is not Japanese at all. Nevertheless, it does remind one of sushi (cold wet dead fish).

Google DIY/Linux Market Share

The fact that Sun overtook Dell in the server market throws my mind into complete confusion when I read that Google is the world's 4th largest maker of computer servers. The story on Manageability.com dates back to October 2004, so it could be that Google is now actually third (still a little ahead of Sun, and Dell has now stumbled).

Google has lots of hardware. In October 2004 it had 450,000 servers—not so much a server farm as a server empire—spread over 25 or so locations around the world. The neat thing is that most of them run “one application” in a parallel kind of way. No point in Google buying expensive anything. Better for it to buy commodity servers for the sake of cheap CPUs and create a grid. Better still, buy the boards direct from AMD and hire a few guys with screwdrivers.

Does Google's DIY approach mess up the server OS figures?

If Google is 3 or 4 or so in server manufacture and all its servers run Linux (and I believe they do), then that's probably 10–15% of the server market running Linux just by counting Google. I wonder whether IDC counts these figures in? I think not. It never lists Google as a server vendor, for obvious reasons (it isn't one) and it's the server vendors that the stats are compiled from.

If this is the case then the Linux server market share figures that are often quoted are wrong by at least 10 percent—and all other figures for UNIX, Windows, OS390, etc are also overstated.

SeeWhy ESP

I ran into yet another entrant into the ESP market last week. The company was founded by Charles Nicholls, who used to be an executive of Business Objects. For me, SeeWhy is an interesting company. While the high profile ESP companies (StreamBase and Apama) have gone after the attractive (and perhaps lucrative) bank trading applications, SeeWhy is unashamedly about BI. That's the reason for the name of the company, I guess.

Personally I like the fact that SeeWhy's proposition isn't about who can scale to the largest number of transactions. It's about answering the question “what is real-time BI?”. As regards how fast you can process multiple streams of events, the most crucial technical requirement is that you have an in-memory architecture that doesn't introduce any latency in its management of streams and which performs in-flight calculations. You need parallelism and you need some close-to-the-iron software engineering. Once you have that, the performance distinctions are probably not too important for the majority of ESP applications. (Did you ever hear of a company dominating its field because it chose the best performing relational database?)

So SeeWhy appeals to me because it doesn't go into a big song and dance about the velocity of the architecture. Instead it goes into a song and dance about real-time BI. I had an engaging interchange with Charles about this. What SeeWhy does is provide real-time graphical representation of cycles in business events. The events themselves can be anything; the movement off the shelves of supermarket goods, the delivery of shipments, the number of mobile phone calls being made and so on. All such events tend to pulse in a cyclical way. There are daily cycles, weekly cycles, monthly cycles and annual cycles—and others probably.

What SeeWhy does is set a baseline (using historical records) and report on any variations from the norm. It can do this at an atomic level (such as a specific supermarket shelf in a specific location) or an aggregated level (several supermarkets in a given area). It can also be programmed to generate alerts (to email, SMS or whatever) when a variation goes beyond a specific threshold and you can have the trend graphs on a screen and drill down if the mood takes you.

If you haven't realised, this is very different to the reports and dashboards of traditional BI from Business Objects, Cognos, Oracle et al. I'm thinking that's it's the beginning of real-time BI.

AVID: An Ex-Technology

AVID, as regular readers of this blog know stands for Anti-Virus is Deceased or, alternatively, AV is Dead. Borrowing from Monty Python, we can emphasise this message by insisting that;

AV's passed on! This technology is no more! It has ceased to be! It's expired and gone to meet its maker! It's a stiff! Bereft of life, it rests in peace! If you hadn't nailed it to the PC it'd be pushing up the daisies! Its metabolic processes are now 'istory! It's off the twig! It's kicked the bucket, it's shuffled off its mortal coil, run down the curtain and joined the bleedin' choir invisibile!! THIS IS AN EX-TECHNOLOGY!!

I'm getting ahead of myself, perhaps, but there's two reasons why we're predicting the imminent demise of AV software.

Point 1; it doesn't work. To be precise it fails to stop the potentially most damaging viruses—the new ones.

How bad is it at doing this? Pretty awful really. Let's take a specific case as an example; the SQL Slammer worm—which first saw the light of day in January 2003. It was estimated that the SQL Slammer worm infected 90 percent of the computers that it could infect in the space of 10 minutes. From the perspective of AV software, it was disastrous. As far as I can tell there is no evidence that any AV software stopped the initial onslaught of this particular worm from the moment it began its jaunt across the Internet.

Why was that?

Well, it was simply a matter of how the worm worked. Once it had infected a computer running Microsoft SQL Server, it scanned the Internet for other such machines to infect and when it found one, it infected it. This caused an explosion of processes looking to infect other machines all running at digital speeds. This little worm didn't need any help from people to proliferate. The way to kill the infection was to download a SQL Server patch which eliminated the buffer overflow that the virus was using to hijack servers. It was the only cure. The cost of SQL Slammer was estimated as $1.5 billion.

Point 2; There are products that actually do the job properly from security vendors Bit9, SecureWave and AppSense.

AV becomes an ex-technology when companies start buying this technology, a trend which is now happily in progress. When this trend explodes, it will be the beginning of the end for viruses and other malware. Until it happens such software will persist because AV technology doesn't stop it effectively.