Chris Webb Interview with Nigel Pendse

Mark Rittman -

If you've got a spare twenty-five minutes, Chris Webb has an interesting podcast featuring Nigel Pendse up on his website at the moment. Chris is a kind of alternate version of me who focuses on Microsoft's OLAP Products (a younger, more successful version probably) and he's been blogging recently about the new features coming up with SQL Server 2005. Nigel Pendse is of course someone who I've got a lot of time for, he's the guy behind the OLAP Report and the OLAP Survey, I've mentioned him a couple of times before and he helped out with my Oracle OLAP vs. MS Analysis Services article for DBAZine.

Anyway, if like a lot of Oracle BI and OLAP consultants you've been awaiting SQL Server 2005 with a bottle of whisky and a loaded revolver pointed at your head you might find the podcast quite interesting. Given that Nigel's been quite a big advocate of Analysis Services in the past, and the podcast was on the effect that SQL Server 2005 would have on the BI market place, I was expecting Nigel to pretty much say that it'll be game over for the other vendors. It doesn't quite turn out like that though, and whilst by no means is Nigel saying that SQL Server 2005 won't go down a storm and that it's not an excellent product, there are quite a few reasons why Microsoft won't have the BI market all to itself.

I won't spoil it by revealing too much but there's a couple of points that Nigel brings up that are worth noting here.

  • The primary driver for sales of SQL Server 2000 has been its price - even more so for Analysis Services as it's free with SQL Server (contrast this to $20k per processor for Oracle OLAP, on top of $40k for Oracle Enterprise Edition). With 2005 however the price of SQL Server Enterprise Edition is 25% higher, which will probably mean considerably less take-up than if prices had stayed constant - and unlike Oracle you pay for the whole thing again to get the upgrade, you don't get upgrades for "free" (or at least as part of the support cost)
  • The fact that BI development for SQL Server 2005 is now done using Visual Studio might well scare off business users who were otherwise happy with the SQL Server 2000 cube designer. From Oracle's perspective they've probably got the more user-friendly cube development environment with Analytic Workspace Manager now, although AWM is a much more primitive environment compared to Visual Studio and the new BI Architect product.
  • Nigel's view is that SQL Server 2005 is unlikely to attract any more finance users than currently use SQL Server 2000 because they prefer to use products such as TM/1, Hyperion etc which have better Excel interfaces and faster/more flexible cube writeback features. Nothing for Oracle to get too excited about here as of course the Oracle OLAP add-in doesn't even feature write-back yet, and of course what we're referring to here is Microsoft's own Excel interface - the ones provided by the many MS AS third-party vendors are of course often excellent. Nigel's view is that MS AS 2005 is going to particularly appeal to application vendors, consultancies and so on, it's unlikely to gain much additional traction with regular business users.
  • Here's the interesting one - the thing that Microsoft's line up still lacks is a coherent BI product line. You've got SQL Server and Analysis Services - the server backend - and Office (Excel) is the query tool. Sharepoint is the portal element and Reporting Services is the report server. Now arguably the OLAP server and Reporting Services are better than their Oracle equivalent, and pricewise you can't beat them, but as products they're on different release cycles, and for Office the business intelligence element is only a small part. Vendors such as Cognos, Business Objects, and Oracle, have put together integrated suites with co-ordinated releases, single price points, integrated security and so on, and there's no such line-up available from Microsoft. Also, vendors such as Business Objects and Cognos (and to a lesser extent, Oracle) have invested a lot of time and effort in the tools and infrastructure needed to support large user communities - ways of scoping data for sets of users, managing data and report permissions, all the paraphernalia you need for lots of data and lots of users - for which there's no Microsoft equivalent. You can certainly get these abilities by buying products by vendors such as ProClarity, Panorama and so on, but then you've got a multi-vendor solution and that's what customers are trying to get away from. A bit of credit to Oracle there for their foresight in this area.

Anyway, if you get a chance have a listen to the podcast. There's a lot good about SQL Server 2005 and I guess with Chris and Nigel speaking, a lot of this is kind of assumed, and what we're looking at here is really the other side of the coin. Top work from Chris, some very erudite advice from Nigel and I'll have to do something myself like this in the future.