Ok – once the marketing panache settles (view it as good, as bad, as fun, as evil – your call – but it does what marketing does, and does it well) – what’s the TECH behind what was announced today?
There are 4 big parts:
- Part 1 – Next Generation VNX.
- Part 2 – Important “Halo” releases around VNX – Appsync, XtremSF
- Part 3 – ViPR General Availability
- Part 4 – “Project Nile”
At the Speed2Lead launch – the other big surprise was “Project Nile”. Here’s what prototypes look like.
What the @#$%? What **IS** this thing?
Techo’s like me tend to underestimate the importance of not only the tech – but the customer consumption model and value model.
VMAX Cloud Edition has been a huge hit. It was our first crack at:
- “maybe more and more customers and service providers want pure utility consumption models – with total up and down flexibility, and no commitments…”
- “maybe the days of provisioning storage are changing to a model of simple consumption vs. provisioning…”
- “maybe everything will be provided as a storage policy over time, and consumed via a multi-tenant portal idea…”.
Our biggest internal problem is that our internal business systems were never built for this (they were built for “build, book, ship”) – and we need to re-tool as this new model continues to scale.
Now – before everyone over-rotates – the reality is that for MOST enterprise and SP customers – the traditional “storage is capex” model is actually a better TCO.
Remember that people tend to be polemic. They want to argue: “It’s THIS WAY! No, it’s THIS WAY! (idiot!)” :-) Reality is always more nuanced, and the “future is here, it’s just not evenly distributed”.
For most workloads, people will take time to move away from traditional “this goes there” models. But – VMAX Cloud Edition has proven that those ideas are RIGHT.
VMAX Cloud Edition uses a pre-cursor of a ViPR controller idea as the “abstractor”, and VMAX under the covers (will move to ViPR as ViPR GAs and matures).
But the more IMPORTANT point of VMAX Cloud Edition isn’t “how is it constructed”, but “how is it consumed”. It’s a value prop, not about the tech. The VMAX Cloud Edition model is that when you buy “20TB of Bronze + 30TB of Silver + 10TB of Gold”, without knowing what you ordered (not unlike public cloud), you get a single engine VMAX, configured with the requisite capacity and IOps. As you scale up, EMC installs more engines and capacity. Some customers pick a capex model. Some chose partial utility models. Many customers pick an economic model where EMC carries the capital, and it’s a pure utility.
BTW – the list price of VMAX Cloud Edition Bronze is around $3.60/GB – ALL IN (maintenance, and assuming a 3 year TCO period – and pricing does vary based on the utility model the customer picks). WHOA. That’s effectively lower than AWS EBS. Sure, Silver, Gold, Platinum have more price – but almost across the band, it’s lower cost effective that EBS (in fact, that become MORE true when you have an IOps centric load).
“Project Nile” (named after a big river that has less dangerous fish than another big river – wink wink) takes the learnings from VMAX Cloud Edition – and applies it more generally.
- Now… Imagine Isilon (huge volume NAS and with ViPR object models), 1-100’s of VNX (small, super dense IOps and transactional GBs), Scale IO on COTS hardware, “cold storage backup tier” (Data Domain) [UPDATE: TO BE VERY CLEAR – THESE MORE “$/GB” vs. “$/IOps” IS WHERE WE THINK THE MAIN TARGET LIES] – but could potentially include XtremIO (insane IOps density for dedupable workloads), and VMAX (incredible reliability/serviceability/availability), a – all abstracted behind a cluster of ViPR controllers. Each of those can smoke the economic models of various existing public cloud models.
- Now… Imagine all that is a simple service catalog.
- Now… Imagine that it wasn’t a pile of TECH, but rather an economic storage utility.
- Now… Imagine that it has simple consumption model.
- Now… Imagine that any enterprise of service provider could get it.
It will take time for us to keep baking this idea. It’s a “Sea of Storage”, “Huge Data Lake”, and “Enterprise elastic transactional storage”. But just like ViPR – we talked about it, and damn the torpedoes we will deliver it. On stage, Jeremy said we are working to a early 2014 target.
THAT. IS. BIG.
Comments always welcome!