Virtualisation - is there an alternative?

Another thought about virtualisation - following my last blog, I started to think about mainframes. Should the mainframe be an option when considering virtualisation projects?

I'd argue that the mainframe does have a role in the modern enterprise. It was there first in so many ways: high performance transactional computing, virtualisation, multi-tasking, multi-OS - I could go on but you get the point.

Though being there first is not to argue that a technology should continue forever, of course. Look at DEC's mini-computers: where are they now?

Rather, enterprises should consider the mainframe as an option when thinking about virtualisation projects.

There are many reasons why that's so: IBM is fully committed to the mainframe, there's a growing and enthusiastic community, and the platform's capabilities have been growing for years.

The main problem for the mainframe is that expertise is starting to dry up, as administrators age and start to consider retirement. However, IBM has made moves to kick-start educational activity both outside and inside academic institutions, so this issue should fade over time. Additionally, IBM and others are developing tools to help cut the management workload.

Just put the mainframe on the table (as it were) and seriously look through the numbers over a realistic timeframe. Instead of re-creating the mainframe by developing orchestrated data centres, why not just get one that IBM prepared earlier?