Steve goes Down Under!

I had an informal invite from Andre Leibovici to chat with the APAC Virtualization Community on his regular Talkshoe and really enjoyed the questions and opportunity to give a view from a front-line trench monkey like myself.  I’m not an official UCS spokesperson so it was a freestyle event that hopefully gave the attendees some unique value :-)

Check out this APAC event and previous events

What did we talk about?  Lots! Here’s the freestyle agenda that was unplanned, live and uncut!

  • What’s UCS all about - how UCS is not about blades, but about virtualizing everything: the whole enchilada!  Nobody else is doing this, only Cisco, end to end.  It’s strategic.
  • What’s the market - 2009 was about early adopting, medium sized agile organizations.  2010 is about more of those organisations and adding the large service providers and enterprises building their *aaS offerings.  The U will soon mean Ubiquitous.
  • What are customers saying – UCS is different in a good way, it makes our lives better.  We don’t want to go back to the old ways. It takes little effort to manage.
  • What’s the differentiator from competition – it’s not about blades, or CPU, or memory.  It’s about scaling at lower cost, enterprise management of your choice (pick best of breed), and wire once and walk away.  Cisco have designed this from scratch, no legacy.  The competitors are just chucking bits into a bag and saying “me too”.
  • What Service Profiles do – over commit compute to drastically reduce CapEx for ALL WORKLOADS, no matter what the hypervisor or native enterprise application.  And reduce OpEx by shorter release processes and workload lifetimes.
  • Eggs in one basket – with powerful Intel chips and massive memory, isn’t it risky to run hundreds of VMs per blade?  If Risk = Probability and Impact, here’s where UCS helps: Probability is reduced because less hardware to fail and less admin involvement which is 50%+ cause of downtime.  Impact can be mitigated by vSphere HA, redundant compute nodes and other resilience techniques.
  • Four sockets coming soon? On C series yes, but B series blades sweetspot is 2 sockets and more cores.  This is sweet because of drastrically reduced licence cost for same Oomph, like 30%+ reduction in Oracle licensing on UCS B200 compared to a 4 Socket AMD – that kind of saving funds a UCS system on its own!
  • Palo / VIC – what I’m excited by is not the networking aspect, or the vCenter integration, but running ANY workload on SAME blade.
  • What workloads – ANY hypervisor, though vSphere is biggest partner and most popular workload today.  Microsoft are doing an awesome job. Customers are using UCS to “virtualize” the 80% that vSphere isn’t.  KVM is rising fast.
  • What management – BMC are stepping up, we’ve just seen what an awesome offering they have.  newScale are announcing UCS developments today.  IBM Tivoli is being used by enterprise to monitor and manage.  CA are on their way.

No related posts.