Home | Virtualisation | Reflections on VMworld 2013
Reflections on VMworld 2013

Reflections on VMworld 2013

3 Flares Twitter 0 Facebook 1 Google+ 0 StumbleUpon 0 Buffer 2 LinkedIn 0 Filament.io 3 Flares ×

Although I haven’t attended this week, I have been keeping more than a passing eye on the events unfolding at this year’s VMworld 2013.  After having absorbed some of the keynotes, blog posts and tweets, it’s about time I expressed an opinion on some of the announcements so far.

The Software Defined Mantra

Say it often enough and people will come to believe it.  We now live in a Software Defined Data Centre world, where everything is in software.  Not quite.  Hardware still has to exist somewhere, however we’re certainly in a place where all of the physical components of compute, networking and storage can be abstracted away from direct access.  In essence that’s what we’ve been doing for years – look back at the original hypervisor, IBM’s VM, or the rationale of any operating system.  However being slightly less cynical, in the x86 world, we are now at the point where we should be able to deploy infrastructure in hardware slices of our choice and let the software to the rest.


I’ve never really got networking.  It seems such as simple idea to me – push packets from one place to another, however invariably, it seems to have been over-engineered in so many places I’ve worked in.  No situation more exemplifies this than when the networking team are allowed to implement Fibre Channel infrastructure.  So, virtualising the network seems like a great idea, especially as a huge amount of the configuration of the network must be for virtual infrastructure anyway (I hasten not to say “traffic” as there’s a difference between configuration overhead and the traffic that flows across that configuration).  Whilst I can’t comment on the virtues of VMware’s design, what I do wonder is how VMware’s implementation of virtual networking will fit into a wider architecture.  Will NSX run separately from ESX?  Could I for instance buy virtual networking from VMware and take my compute from Microsoft with Hyper-V?  I hope you will be able to, but alas suspect not (please correct me if I’m wrong though).  Keep that thought for a moment.

vSAN, vVol, Flash (Read) Cache & Virsto

VMware virtual storage – quite the collection now.  New virtual SANs (is this VSA revamped?), vVols (although not yet announced or available), Flash Read Cache and their acquisition, Virsto.  With EMC as the parent company, what’s an architect to think?  Should I be using external storage any more?  Which is better, Virsto or vSAN?  Will both of these support vVOLs or will that be purely for external arrays?  The message is confused and is made even more confusing with EMC in the mix and their ScaleIO acquisition.  I think the sales team will have a difficult time positioning each of these technologies, however I will have a go.

  • Virsto – VDI competition beater to try and kill off SSD arrays for VDI (hello Violin & frenemies)
  • Virtual SAN – noSAN competition beater to try and kill off Nutanix
  • vVOLs – intelligent array beater to try and kill off Tintri
  • Flash Read Cache – take your choice here – Proximal Data, PernixData, plus a range of PCIe SSD vendors

OK, that’s the cynical part over with.  In reality VMware will always see other startup features they can either acquire or emulate and integrate into their platform.  Storage isn’t immune from being SDx’d although the issues of dealing with persistent data are different.  But, here’s the thing to think over; how many of these infrastructure components will be compatible with other hypervisors?  The answer again is none.

The Architect’s View

VMware are doing a great job of evolving the data centre and the move to Software Defined is a good thing.  vVOLs are definitely a step forward; Virsto was a great acquisition.  However if you’re not all-in with VMware, then getting the most out of their new features will be tough.  VMware are a business and it’s all about control and market share.  They want you using their technology and theirs alone.  Although this isn’t a bad thing in itself (and something I’m not naive enough to expect all vendors to do), it’s one angle that needs to be thought through.  If you’re happy being fully VMware integrated, then that’s great.

On a positive note, I like the sound of the new features listed above and I look forward to trying them out and seeing how they fit into an overall virtualisation strategy, so ultimately it’s not all bad!

P.S. – I realise I’ve skipped over quite a few other announcements.  They will have to wait for another day.

Comments are always welcome; please indicate if you work for a vendor as it’s only fair.  If you have any related links of interest, please feel free to add them as a comment for consideration.

Subscribe to the newsletter! – simply follow this link and enter your basic details (email addresses not shared with any other site).

Copyright (c) 2013 – Brookend Ltd, first published on http://architecting.it, do not reproduce without permission.

About Chris M Evans

Chris M Evans has worked in the technology industry since 1987, starting as a systems programmer on the IBM mainframe platform, while retaining an interest in storage. After working abroad, he co-founded an Internet-based music distribution company during the .com era, returning to consultancy in the new millennium. In 2009 Chris co-founded Langton Blue Ltd (www.langtonblue.com), a boutique consultancy firm focused on delivering business benefit through efficient technology deployments. Chris writes a popular blog at http://blog.architecting.it, attends many conferences and invitation-only events and can be found providing regular industry contributions through Twitter (@chrismevans) and other social media outlets.
  • http://etherealmind.com Etherealmind

    Just to respond on the networking issues.

    Yes, VMware NSX runs a separate controller implementation. I think this is because it’s hypervisor independent and works for KVM, ESX and Hyper-V. There are differences in implementation but NSX works for the common hypervisors as a single coherent service.

    Your attitude to the network fails to embrace the concept of network services as part of the network. I guess that a network to the storage industry is just a fibrechannel box that has some ports in it that is performing cable emulation, and therefore you expect little from your networking equipment. However, Fibrechannel is remarkably inefficient protocol that is unable to scale well or perform in large diameter networks.

    The storage industry has become inured to the high costs of their legacy systems and doesn’t tend to adopt new technologies, so perhaps the pace of change is a bit too exciting for some.

    Networks services include firewalls, IDS, load balancing, accounting and inspection. And VMware NSX (and, I believe, so will Cisco’s SDN response) does all of these things.

    In relation to VSAN, this will have a significant impact on network designs in the future. The current limitations on size of the just eight machines means we have some time to consider what those changes mean.

  • Jeff Frick

    Watch theCUBE Interviews here


  • Pingback: Virtualisation – Solving the Storage Problem | Architecting IT()

3 Flares Twitter 0 Facebook 1 Google+ 0 StumbleUpon 0 Buffer 2 LinkedIn 0 Filament.io 3 Flares ×