>>>But in a prior message, he referenced this in conjunction with SQL Server and some of their decisions on functionality in the DBMS. Maybe I misunderstood, as several things were being talked about. I'm not trying to make a particular point - I truly didn't understand what he was saying.
>
>Because I didn't say it. The context is autospanning: how does it apply on a smartphone with 32G memory? Seems to me that the smartphone is the Ford Model T and autospanning is the leather bridle that I used to swear by. But now? Whether manufacturers put leather bridles onto their vehicles is of little interest. To me, at least.

And by the way, on the in-memory vs autospanning debate, it isn't "dead", it's just taken on a different form

SSAS OLAP is a tried and true methodology in Microsoft SQL Server. You can have cubes and OLAP partitions that greatly exceed memory space, so there is something akin to autospanning under the hood.

There is also a new SSAS Tabular Model - it compresses data and is purely in-memory. The compression is great, but if the size of the tabular database (compressed) exceeds server memory, you're out of luck.

And that's why I've stayed with the traditional OLAP/MOLAP model. See the irony? :)