Reaction to this morning’s VMworld Europe keynote are coming in.
This is powerful stuff. The offline VDI stuff really enables an
entirely new way of thinking about VDI; it’s no longer about just
hosting desktops at the datacenter. Now it’s about providing a “golden
image” that users can run on the local machine when they’re not in the
office and on the server farm when they are in the office.
Lode Vermeiren on his "vmworld" Twitter channel during the keynote had this to say about our application virtualization technology (aka Thinstall): "Mark my words: This is fan-tas-tic technology!"
Adrian Bridgwater, blogging at ZDNet.co.uk, took notes on how Diane described the maturation of this virtualization. (We’ve been working on this for 10 years, and we think it shows.) Link: The evolution of virtualisation
1 – SEPARATION: the process of being able to test for all possible system configurations before deployment.
2 – CONSOLIDATION: after separation, the move to server consolidation can happen.
3 – AGGREGATION: this is the point at which a virtual data centre is
built to run services as per the requirements of the system. (British
Telecom was present to give a short presentation on its work building
11 data centres where it has apparently seen a 50% reduction in the
cost of physical servers.)
4 – AUTOMATION: this is the day-to-day operation of a self-managing data centre.
… and finally (and this is where we’re at today).
5 – LIBERATION: this is where data is utilised from the ‘greenest
available resource’ and workloads are moved around (globally) to make
sure this happens on the biggest and best scale possible. In Greene’s
own words this type of technology is an, “Early step towards cloud
[Update: we’ll come back to desktops later, but I also wanted to call attention to this from Michel Roth: Link: Scalable Virtual Image – Thincomputing.net.
A relatively small engineering effort from VMware yielded an
announcement at VMworld Europe that could have a severe impact on VDI
implementations: the Scalable Virtual Image.
Why do I say "relatively small engineering effort"? Well, because
it’s based on the existing "linked clone" technology that you probably
already use today in VMware Workstation (6 and onwards).
Why do I say "could have a severe impact on VDI implementations"? Well,
because as of today, one of the drawbacks of implementation VDI still
is the storage requirement. Sure, there are lots of companies out there
that are jumping in to fill this void (streaming, other cloning
technologies) but the fact that these are other companies also implies additional costs.
Anyway, the technology is not available today but this is very interesting stuff to watch