I'm not a bit BI (Business Intelligence) user. I think the promise is great, and I was excited when SQL Server gained an OLAP component in SQL Server 7. There have been many changes over the years, but overall the investment in Analysis Services (SSAS) has tailed off in recent versions. In fact, there has been speculation that Microsoft will abandon SSAS in the future and remove it from the SQL Server platform. The few changes in 2014, more, but still nothing major in 2016, has had users worried about continuing to build applications (and their skills) on SSAS.

I'm confident that's not the case now. This past week Microsoft announced Azure Analysis Services at the PASS Summit. This is the evolution of SSAS on the SQL Server platform, with the ability to now move your tabular models into Azure and run them on an as needed basis. This means that you don't need to administer your own SSAS instance, and can connect to cloud and on-premise data sources for your data analysis needs. Since SQL Server has moved its codebase to primary development in Azure and a periodic release on-premise, this is a good sign that Analysis Services will continue to receive investment in the future.

There are restrictions. No multidimensional models, for one, and a few others. If you are interested in trying out the SSAS in Azure, make sure you understand that limitations. And since this is in preview and has been announced, keep up to date on the changes in the platform. As developers build in new features, I'd expect this platform to evolve and change at least quarterly, with new features and fewer restrictions over time.

I know others are excited, as I've already seen a few people blogging this week about their initial experiences. As with other services in Azure, however, I both like and dislike some things. It's great that the platform evolves and changes quickly, but I'd like to know which release of Azure Analysis Services I'm using. If for no other reason than I'd like some idea of when something changes. Not every evolution is helpful, and some will break systems, so knowing there has been a change or a new release can dramatically speed up troubleshooting. If the version of the Azure system changes, then I know to potentially look at Azure release notes rather than my own code. If it doesn't, then problems are more likely my own fault.

It hasn't seemed that SSAS has grown as much in use or popularity than I had expected 17 years ago, and I suspected Microsoft might give up, but that isn't the case. For now, we see that SSAS is alive and well, moving into the cloud and receiving development resources. Perhaps that will change, but I hope not. There are problem domains solved well by SSAS and the ability to use the technology as an on-demand platform, without adding additional administrative and hardware resources, might help grow adoption a bit. I still think the core understanding of SSAS and MDX escapes most people, but we have many more tools that can read data from SSAS instances and help users query data. That means we need fewer people to design and maintain SSAS instances, and we can easily create and destroy them as needed on Azure.