By Adam Hawley on Aug 23, 2011
Let me hit you with some quick highlights for Oracle VM 3.0, announced today:
- All the management you really need to support your entire data center: Network management, storage management, dynamic resource scheduling (triggered on network not just CPU!)
- An even more scalable server: Support for 128 vCPUs per guest VM vs. 32 vCPUs for VMware.
- A more scalable budget (your's): Save about $2.5million dollars over 3yrs vs. VMware (license + support. See configuration details below but their pricing is here). Would you like to have that in your pocket right now? Or maybe just have it for one of these?
And that's just the beginning. The definition of what represents "value" in a virtualization solution just changed. Oracle has been focused on delivering an increased feature set at the same price ($0 license, pay only for support) while others have been focused on delivering increased pricing (though perhaps "focus" is not the right word here).
Since the day we created Oracle VM, we've focused on exceptional price:performance and scalability. That focus continues, hence the support for 128 vCPUs and up to 160 physical CPUs, but with Oracle VM 3.0, we're introducing a massive set of enhancements to manageability and ease-of-use that will transform the way you can think about your options for deploying Oracle VM: Now you get not-only exceptional price:performance but also all the management features you really need to deploy Oracle VM across your entire data center and not just for your Oracle software. In the past, Oracle VM required some configuration from the manager and some configuration at each individual server, which meant a little more work to get going and to maintain things in very large environments. For some customers, this meant they mainly focused on doing that only for their Oracle applications and directly related things, but not much beyond that.
But no longer.
With Oracle VM 3.0, everything is done centrally from the management server: discovery, server-, storage-, and network set-up all from a browser (no client software required, no logging in to each individual servers). And then, once you've got it set-up, it has a rich feature set for managing and maintaining your environment:
- Dynamic resource scheduling to load balance across server pools based on network (and CPU): trigger live migration (migration of running VMs) based on CPU and/or network load. Note that I think triggering DRS on network load is a first amongst the major vendors anyway. This capability combined with one of our network management features that allows the separation of different traffic types onto different, independent networks (e.g. separate guest IP traffic from storage traffic and others) means that you can trigger live migration to occur over a dedicated live migration network based on hitting a load threshold on the guest IP network. This separation is important because if you start a live migration over a shared network that is already overloaded, you risk making things worse. But we make it better instead. Very important for enterprise loads that are very I/O intensive
- Dynamic Server Power Management: automatically power-off underutilized servers without disrupting your applications; power them back-on when needed.
- Comprehensive Event Logging: Event logging sometimes sounds like something pretty boring, but its not. It's really not. Just ask any admin trying to track down a change in the environment in the middle of the night whether he or she thinks they'd like to have access to event logs that can tell you basically anything that triggered, who did it, and whether it succeeded or failed and why and I'm very sure they'd say "yes please". And we put this all nicely available centrally in the UI in calendar form: click a date on the calendar to see what happened in the environment instead of having to dig through a bunch of obscure logs that you need to "tail" to try to understand what happened.
There is so, SO much in 3.0 I could write a massive entry here going on-and-on but, instead, the best thing to do is to watch the replay of the live event (once posted...it is being processed today for replay but should be on the site here any minute), including the tech webcasts featuring myself and my team that focus on different aspects of the product.
Oh yeah...the config that will cost you an extra $2.5 million from VMware is for 250 2-socket servers, with 96GB of "vRAM" per server on Oracle VM 3.0 vs. VMware vSphere5 (VMware's revised pricing). You can compare any configuration you want by trying out our cost-calculator here.
In the coming days and weeks, keep an eye on this blog. We'll be focusing on entries that go into the details in the various feature sets of Oracle VM 3.0. But the proof of the pudding is in the tasting as they say so the best test of all is to get the software and try it out for yourself. The software is available today for existing customers through My Oracle Support. For users without a My Oracle Support account, the software is expected to be available on Oracle's eDelivery website in the next 24hrs.