3 thoughts on “Storage Smackdown Brownbag (12/7) – A Call For Questions!”

1. I’m more curious about the roughly 25% utlization on the VASA. At that price point you’re not getting very much in the way of performance or space.

   Reply ↓

2. Do you see large enterprises adopting unified platforms such as FlexPod or Vblock in place of their deeply-entrenched and traditional stack infrastructures? Or is it more of a niche play at the moment?

   A: Most customers are still doing “mix and match” rather than buying converged infrastructure (bought integrated, managed integrated, supported integrated). Usually, it’s driven by someone OTHER than the infrastructure team (their leadership usually), who realizes how little value the enterprise gets out of “putting together pieces”, even when there is a well documented reference architecture (ahem). That said, it’s picking up. We’ve publicly disclosed that Vblock booked $140M of revenue in Q3 ’11, and that the number of net new customers is ramping like gangbusters, and that the average repeat order (a good testament of value) is around 38%. I would expect that over time, this model of IT consumption will become more an more prevalent, which is why Cisco and EMC (with VMware and Intel) formed VCE – a joint venture, because clearly while Cisco and EMC could sell (through our channel partners) VMware solutions, on Cisco and on EMC (I call this “V+C+E Reference Architectures”), if a customer wanted integrated infrastructure – we needed to do something different.

   How do you envision the evolution of technologies like intelligent caching algorithms (VST, FAST) and server-side SSD memory caches (Project Lightning) shape how customers consume data?

   A: Customers, in the end, don’t care What they want is the best performance at the lowest cost – when performance is the primary measure, and most cost effective at the lowest cost. Ultimately – they want both. Right now, host-side Flash has narrow use cases, but in those it just rocks. Right now, array based SSD (and Flash as Cache models) has broad use cases (and in those it just rocks). With the continued commoditization of solid state memory – this will just continue. Where we think a lot of innovation will be around linking these ideas. Flash will exist everywhere. There will come a time where every server will ship with Flash – so you can imagine that transparency (and persistence!) of data location will need to become automated.

   As geographically distributed storage solutions like MetroCluster and VPLEX become more popular in practice, what can you divulge about the future innovations of continuous availability solutions?

   A: Expect to see Async solutions (we already do this with VPLEX Geo, but in narrow, narrow use cases) to come alive. Expect to see the network complexity be reduced by things like VXLAN. Expect to see pure software implementations (VPLEX is really just software). Expect to see more “hardening” of vSphere in stretched clusters (next little – but big in practice – thing will be host HA response on PDL (look it up Expect to see us be able to “push” a VPLEX instance to a cloud provider when you need it (we actually demoed this prototype at EMC world). Expect to see vSphere stretched clusters and SRM become an “and” not an “or”. Much, much more! This space is a little rare today, but will be mainstream tomorrow.

   How do you see VMware’s expansion into providing storage technologies like the vSphere Storage Appliance (VSA) and host-based replication compliment the current SMB storage array offerings?

   A) from an EMC perspective – it’s awesome! Customers choose EMC SMB solutions (like Iomega PX for a few thousand or a VNXe for less than $10K) or VMware VSA – it’s all good the VSA starts below the band of a VNXe – if a customer needs more performance (about 10x more), or more scalability (about 10x more), but wants all the simplicity and ease of use (including vCenter integrated management), we offer a great little upgrade – and easy to move Oh – and the list price is about the same. That said, there are use cases where pure VSA approaches make sense – sometimes even the small 2U of a VNXe is too much. One interesting thing is EMC studied a lot of customers to see whether they wanted 3rd party VSAs (imagine a VNX as a VSA), and in general, they were too confused around the “inception like nature” of the model – there was a strong preference for it to be offered as a native VMware feature – so, happy to see it offered that way Likewise vSphere Replication 1.0 is free when you use SRM5, we tend to see it used in mid-to-large customer in conjunction with array-based replicas, but it’s awesome for SMB customers, or folks for whom it’s not an option for one reason or another. We’re also pretty pumped over what it will enable in terms of “DR to the public cloud”.

   Has there been any progress made around the recent performance issues with the VAAI Thin Provision Block Reclaim primitive (UNMAP)?

   A: YES. We’re getting close. Expect new patches to vSphere to decouple svMotion completion from the delete (making it async) – this will be the near term fix. It’s not out yet, FYI – so check out the workaround to disable UNMAP. In addition, we’re working on making the UNMAP commands be much more async on the array side too.

   Thanks for the questions!

   Reply ↓

3. Pingback: BrownBag Follow-Up – Storage Smackdown 2011