Last week I have been to VMWare’s yearly conference VMWorld in Las Vegas. After last years attendence record of 17.000 this year it was estimated to go up to 25.000 – however thanks to hurricane Irene “only” 20.000 VMWare customers made it to this years conference.
However, that number is huge and compared to the 3000 people that came to VMWorld 2005 in Las Vegas this shows the development of this technology and that VM technology on industry standard hardware has reached the mainstream and a high level of maturity.
So what is new this year? Well, last year was all about the cloud and surprisingly – even though the conferences’ motto was “it is your cloud” – the topic was much more deemphasized compared to last year.
The focus was definitely on business critical applications this year. Interesting slide in the opening keynote which showed that today >60% of business critical Microsoft Application deployment (SQL, Sharepoint, Exchange) is running on VMWare and more than 40% of Oracle and SAP deployments are on VMWare.
From a technical point VMWare now supports 32 virtual CPU’s, 1TB of RAM and up to 1.000.000 IOPS in a virtual machine. That is pretty big! If we take a close look at our own system landscape we will not find anything that would not fit into this kind of Monster VM.
Interestingly enough though – also this was not heralded as the biggest new feature. VM’s have become part of what we do and therefore it seems we come to expect that any system workload can run in a VM anyway.
The most interesting topic was how VMWare will cope with the remaining bottleneck of storage performance. VSphere 5.0 has many interesting features like Storage DRS (moving virtual disks from slow performing LUNs to better performing LUNs or ensuring that growing virtual disks will not run out of space) but also some new concepts in regards to Storage API’s where LUNs as a concept are becoming obsolete and VMWare manages the underlying storage directly and thus eliminating an extra layer of management.
On the management side of things VMWare is pushing VCenter Operations as there self learning monitoring application. It looks impressive in the way it correlates all underlying parameters and learns what behavior of a specific system is normal and at which point an alert has to be generated because the system moves into an “un-normal” stage.
Taken all things together the message is clear – everything can run on VMWare without any exception.
I have to say it was again a rather impressive show – VMWare seems to be very focused on what they do best and where there main business lies but also trying to venture out into new technology (their client approach based on VDI still lacks in some areas but becomes more an more compelling every year).