Kilimanjaro, Projects Gemini and Madison Webcast

For those who haven’t seen much of Project Gemini but have heard the buzz, this TechNet Webcast: An Early look at SQL Server ‘Kilimanjaro’ and project ‘Madison’ – will give you a good insight. It also has some features on reusable Reporting Services components which look very impressive and info on Project Madison, which provides scalability features. Registration as usual is a pain, and forget trying to use the site using any browser other than IE – I wish Microsoft would make their content easier to access.

Anyway, onto the webcast – about the 1st quarter of the webcast is the usual generic roadmap blurb, but then the presenter gets into the real meat of Gemini – an Excel based ‘in memory’ analysis tool that allows joining between entities without having to know about such things, superfast analytics – pivoting, calculation, charts etc. and then being able to publish to Sharepoint. From an OLAP analysis point of view, the Pivot Tables also has slicers (effctively table wide filters) displayed in the spreadsheet as well, and it would be good if that made it into Pivot Tables generally in the next release of Office. It looks like an incredible tool and very easy to use – and may be a powerful step towards the realisation of the ‘BI for the masses’ vision. The presenter did let slip one weakness though – much has been made of the 100 million rows of data demo – but that data still has to be loaded into memory first and will still take significant time. I also suspect that how successful Gemini will be is going to depend on how much it will rely on good data structures being in place in an organisation to support it. The Data Warehouse is going to remain the core part of  any succesful BI delivery.

The next component of interest was the reusable Reporting Services components – there is the concept of a library of components that can be built – e.g. standard charts, logos, gauges etc – and then dropped into any report, either by a developer or a user in Report Builder 2.0.  What really grabbed my attention is that these components are version aware – i.e. if the library version of the component is updated then if you reopen the report in design mode it will let you know and give you the option to update. Again this points to ‘BI for the masses’ as you can have developers create some great components which any user can then drop in to their home grown reports. Plus as any developer knows, there’s a lot of repetition and any options for code re-use are always appreciated.

Finally, Project Madison was covered – and seems more about scalability up to multi-terabyte warehouses. It was a bit infrastructure focused for me so most of it passed me by, but clearly Microsoft are stepping up to try and address the market perception that they can’t scale.

Anyway, this all will be dropped in late 2010 as Kilimanjaro – an interim release of SQL Server.