I've been to an Azure Training day today. The training was great, however it it seems that none of what I learnt was any use to me, or if it was then the costs make it prohibitive.

I was looking at it for scalability. What-Am-I-Doing is likely to need some help scaling. It creates a reasonable amount of data - 600 or so lines of data per day - and I need a way to multiply that by the number of users. Azure is not the answer.

SQL Azure costs either $9.99 per month for a 1GB instance, or $99.99 per month for a 10GB instance. Plus I'd have to pay for data transfer.

A 10GB instance is neither cheap, or any more scalable than a standard SQL server licence.

Rewriting the app to use the Azure storage option is a load of work, and to be honest, I can't be bothered. The reason for using a relational database is so that the data can be easily managed and related.

I think that I get, I just don't get it. There was one person who I could see was going to get great benefit from Azure. They run a public organization that each year runs a public survey. In one weekend they get a massive amount of traffc to collect survey results. This is an ideal use of Azure instances - they could just switch on more instances for the weekend, and then switch them all off when the survey is complete.

It was an interesting day, well enough presented, and the hosts (ICS at Basingstoke) were great. It's just a shame that Azure is not there yet.

They always say that it's not what you know it's who you know.

I was chatting to a friend of a friend who works at Microsoft recently. I was talking about some work I was doing that grabbed data from SalesForce.com and stored it in a SQL server database. the grabbed his attention as he is involved in Microsoft CRM, and his boss is tasked with their compete strategy - helping microsoft sell CRM against other products.

A meeting was arranged so that I could demonstrate my data extraction tool, and then they were going to show me their "Data Migration Manager" (DMM). We'd see whether we could add the two together to create a migration strategy.

Preparation - the 5 Ps.

I was able to download the DMM and a virtual PC image with a version of MS CRM on it, which gave me a bit of a head start. With a bit of playing, I realised that the DMM used XML files to store the maps, and as SQL can describe itself pretty well I was able to modify my data extraction too so that it output a set of CSV files compatible with the DMM, and also an XML map.

The Meeting

The meeting went a bit like this. There was a fair amount of talk getting things started and positionsing everything, and then they asked me to show the extraction process.

It ran pretty smoothly, extracted all the data, and I was able to show it to them in query analyser.

We talked about getting the data out into CSVflils, and I pressed a button and said "you mean like this?" They seems pleased.

We then talked about DMM maps for the CSV files, and so I pressed a button and said "do you mean like this?" They seemed almost excited.

They then talked about the DMM process, and while they were doing so, I used the DMM to load the small test data set into their demo CRM instance.

Follow Up

Obviously, life is never that simple, and there are some compleixties that need to be sorted out - the fact that some relationships are more conveluted and that the source data structure may not match the destination, however we do have the basis for a product.