Microsoft Fast on Linux and Unix Innovation

February 15, 2010

It’s Valentine’s Day. I feel quite a bit of affection for the system professionals who have licensed Fast Search ESP, and I hope each finds search love. I think there will be a “tough” element to this love. And like other types of love, there will be ups and downs. Microsoft practiced some “tough love” for licensees of the Linux and Unix versions of Fast Search & Transfer’s Enterprise Search Platform recently. I am in a discursive frame of mind, and I will share my opinion about the “tough love” for the Linux and Unix licensees of the 1997 technology that comprises some of Fast Search & Transfer’s system.

The not-too-surprising announcement that Microsoft would stop supporting Fast Search & Transfer’s Linux and Unix customers surprised some folks. I think a handful of resellers were delighted because customers with non-Windows versions of Fast Search cannot change horses in the middle of the Tigris River, as Alexander the Great discovered in 331 BCE. Some poobahs pointed out that open source search would become a hot ticket for Fast Search Linux and Unix licensees. Others took a more balanced view of figuring out whether to rip and replace or supplement the aging Fast Search system with one of the more specialized solutions now available; for example, Exalead’s system could be snapped in without much hassle, based on my research for Successful Enterprise Search Management, published by Galatea in the UK last year. (Martin White was my co-author.)

image

Source: http://www.zastavki.com/pictures/1024×768/2008/Saint_Valentines_Day_St.Valentine_004959_.jpg

What I found interesting is that the Microsoft Enterprise Search blog contained some information from Bjørn Olstad, CTO, FAST and Distinguished Engineer, Microsoft. The write up’s title is “Innovation on Linux and Unix,” and it appeared on February 4, 2010.

Mr. Olstad wrote:

When we announced the acquisition two years ago, we said that we were committed to cross-platform innovation—that we’d “continue to offer stand-alone versions of ESP that run on Linux and UNIX,” and that we would provide updates to these versions to address customer concerns and add new features.  Over the last two years, we’ve done just that.

The deal was consummated in April 2008. In October 2008, the Norwegian authorities seized some company information, but there has not been much news about the investigation into the pre-acquisition Fast Search & Transfer’s activities. At any event, it is now February 2010, so Microsoft has been operating Fast Search for the period between April 2008 and February 2010. That’s not quite two years, which is a nit, but software works when details are correct. What’s clear is that Fast Search and its Enterprise Search Platform or ESP is pared down and focused on the Windows platform.

I also noted this passage:

When we announced the acquisition two years ago, we said that we were committed to cross-platform innovation—that we’d “continue to offer stand-alone versions of ESP that run on Linux and UNIX,” and that we would provide updates to these versions to address customer concerns and add new features.  Over the last two years, we’ve done just that.

Another detail. The categorical affirmative “always” does not match with the shift in the direction of the Fast Search & Transfer technology. The “always”, in my opinion, means “10 years.” Based on my experience in enterprise search, I am not certain that the present version of Fast Search for Linux and Unix will have many users in 2020. The core is almost as old as Google’s. Unlike Google, Fast Search has added features by using multiple methods. Google, despite its management methods, has engineered enhancements into the plumbing of Google. As a result, the framework is more cohesive, modular, and in my opinion, more in tune with state of the art search and content processing methods. You may, of course, disagree. In the same league with Google is Exalead because both firms have their knowledge tentacles looped into the learnings of the original Alta Vista team. Frankly Fast Search has not kept pace with either Google or Exalead and I can name a number of other companies who have also outpaced Microsoft Fast in the last 18 months.

image

Image source: http://www.wordseye.com/sl/webpage-db/2006-12-24/7860.jpg

I marked this segment on my hard copy too:

When FAST was founded back in 1997, we were told that it was too late to start a search company. The prevailing wisdom back then was that search was already a commodity: Verity had won the enterprise and AltaVista had won the web.  More than ten years later, it’s clear that we’re just getting started.

I would respectfully submit that the Web search game is Google’s at the present time. Now Google is seeping into the enterprise, and I think Microsoft’s dependence on Windows 7’s revenues underscores that there is a fragility in the Microsoft balance sheet. I am no financial analyst, but if Google opens a vein in the enterprise and in the Windows 7-type market, Microsoft could suffer and suffer in a painful way. I would also point out that organizations have a number of options when it comes to search and content processing. I track more than 300 vendors in my Overflight service, and I think that there are several risks Microsoft and its Fast unit will have to ameliorate:

  1. Giving away a free or low cost search system to SharePoint customers will provide a short term payoff but over the course of a period of time, the costs of stabilizing and scaling Fast ESP may open the door to competitors. Believe me, there are lots of snap ins available for SharePoint. There’s BA-Insight, Coveo, Exalead, Mindbreeze, and more. You can turn to Bitext which has a utility that makes SharePoint search better quickly. And more options will be coming in the months ahead. Proliferating a complex system like Fast ESP will be the equivalent of teaching customers to look for third party solutions to get users happy and control costs. That’s my opinion and you are welcome to disagree. Just bring facts to your push back.
  2. Fast ESP is complex. I have mentioned the 20 page white paper from Microsoft that details the settings that must be configured to deliver results tailored to a specific search implementation. I am not going to rehash that list. Keep in mind that there is a 300 page technical document that I had to work through to deal with some of those Fast ESP settings. Maybe that document is no longer needed. I sure needed it, and I thought I was pretty good at search systems.
  3. Google is targeting Microsoft, but in its own Googley way. My hunch is that as Google pushes out more enterprise solutions, the pressure on Microsoft will increase very gradually. The analogy I use in my client presentations is that of a scuba diver who goes to deep. Everything is great until the diver surfaces and discovers that the wrong air mixture was used. Microsoft is in the euphoric state of having a  new Fast release in the next few months. The resignations and annoyance of the Linux and Unix crowd are essentially irrelevant. But Google, in my opinion, knows the diver’s air mixture is wrong. Google will just let Microsoft Fast ESP happen. Google will be there for those who want to use Google Apps, Maps, whatever.

Please, read the February 4, 2010, blog post. I may be wrong. I hope I am wrong because quite a few information technology professionals have identified themselves closely with Microsoft SharePoint and Microsoft Fast search. If the search system chews through user satisfaction and produces unexpected cost overruns, the consequences could reshape the search landscape. Autonomy, Coveo, Endeca, Exalead, Fabasoft Mindbreeze, Ontolica, and others don’t have to do anything.

Just wait. This time next year it will be interesting to review the enterprise search landscape. By the way, I will be covering enterprise search from my new tie up with Informed Market Intelligence in London. You can find the Web site at http://www.globaletm.com/. More details will appear in Beyond Search when I get the url to the search section to which I will contribute.

Stephen E Arnold, February 14, 2010

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta