James O'Neill's blog

Windows Platform, Virtualization and PowerShell with a little Photography for good measure.

Search: A quick round-up

Search: A quick round-up

  • Comments 2
  • Likes

I've been meaning to have a play with Search Server since Viral demo'd it on the last roadshow. We have a downloadable VHD for it with a very good walk-through, I can say that honestly because I've just spent a couple of hours walking through it.

Of course since I'm doing all my virtualization on Server 2008 and Hyper-V now, I thought I'd try it out on there and rather annoyingly the VHD was built with an out of date set of Virtual Server extensions which won't un-install under Hyper-V.  So I downloaded Virtual PC 2007, which warns me that it is not supported on Server 2008 (and I'm running Virualization on top of Virtualization - a silly thing to do) but it works all the same. So I fired up the VM, removed the extensions, shut it down, booted back in Hyper-V, added the integration components and off it goes. I'd run search server with more than the 1GB specified but even with that it runs OK.

Now I've been playing with Microsoft Search technologies since we introduced Index Server for IIS 2.0 back in about 1996, and at one stage of my life I was a Sharepoint Portal Server Guru. Back in 1999 when I was first interviewing for a job at Microsoft I brought in the some of the work I had done with digital dashboards - stuff that evolved into Web parts. Doing the walk-through for Search Server express two things leap out at me; first was how much easier it is now than I remember it. The other is how simple but how clever federated search is. Some while back I mentioned open search and that seems to me to be a clever technology based on two simple ideas (a) Provide an XML document to tell people how to query your site. (b) return the results as XML to make it easy to consume them in something else. (Use the RSS XML-Schema to make this doubly easy).  Several people are keeping lists of services which support OpenSearch, OpenSearch.org has a list of such lists.

According to Sharon's blog we've booked her to do a series of workshops on search, which will cover search server unlike the walk through I haven't seen the content for these, but knowing it's Sharon doing them, I feel safe recommending those too. Someone is bound to ask her the perennial question "Can I index a database" - answer "What constitutes a match ? What is the URL you're going to click, and what will that open ? - but you can create a web pages to crawl data and return records". I don't know if she's going to cover things like adding PDF support, or whether Viral will have that on his blog sometime soon. I guess if he's going to blog about Virtualization I'll have to blog about configuring iFilters.

Finally this morning I got a mail saying we have Announced an Offer to Acquire the Norweigian search company FAST , I don't know anything about them or how their product will fit in with our other technologies

 

Comments
  • FAST complements Microsoft's search technologies by scaling search beyond 50 million documents which is the current peformance limitation.  

    Search Server Express - no limits but scales to about 1 million docs

    Search Server/SharePoint Search - upto 50 million docs after which scalability is a complex engineering issue.

    FAST - Uber Enterprise

  • Yup, we'll be covering databases - to mash or not to mash..., indexing PDFs and other non-MS content, how to tweak and improve relevancy, will probably touch on information architecture and metadata gotchas, and the social life of information (to quote a book title)... now I just need to free up a bit more than the 900Mb space remaining on my laptop to build the new demo :-)

Your comment has been posted.   Close
Thank you, your comment requires moderation so it may take a while to appear.   Close
Leave a Comment