peshkova - Fotolia

EMC ViPR software-defined storage: Why, and can it succeed?

EMC’s ViPR software-defined storage aims to provide a multi-vendor management layer with cloud and big data capabilities. But can it convince other storage companies to play ball?

Two big questions are posed by EMC’s unveiling of its ViPR big data/cloud/storage virtualisation – or software-defined storage -- platform announced this week at EMC World in Las Vegas.

Namely, why and will it succeed?

Let’s deal with them in that order.

EMC’s justification – covered in more depth here and here – is that ViPR is the company’s answer to a changing storage landscape; one where the customer wants to operate on a private cloud basis; where data from many sources resides on many different platforms, too “heavy” to move, and; that the customer will want to use that data to glean intelligence.

So, ViPR is in large part a storage virtualisation platform. It will knit together disparate storage systems from any vendor (that it has adapters for, and there’s the first rub; there’s only NetApp that’s non-EMC) and from commodity drives. It is also part big data analytics platform that will allow data to be reasoned over in place.

And there’s no doubt it is aimed at a changing future in storage. EMC Powerpointed some IDC predictions to support this, with revenues from existing (ie, enterprise datacentre) IT architectures to grow by 4% annually to 2016 while those from new architectures (mobile, hyperscale, big data) set to grow by 24%, and 78% of customers operating a private cloud in three years’ time.

But at the same time it’s clearly aimed at EMC retaining market share and hooks in customer environments.

Read more:

EMC World 2013: Google storage coming to your data center

EMC aims to define software-defined storage with ViPR

EMC: Data everywhere – ViPR software-defined storage the answer

Raffo: EMC's ViPR 'first step' in defining software-defined storage

Schulz: ViPR aimed at service providers, enterprise storage

Analysts at the event thought so. Gartner research director for storage technologies and strategies, Valdis Filks says: “EMC is preparing for the commoditisation of storage hardware, and as VMAX and VNX margins are threatened it needs to move to stickier software-based systems management. Historically many companies have tried this route, but this time EMC may succeed.”

In other words, while EMC spent a lot of time at its rally of the party faithful declaring how their moves are motivated by giving CHOICE to the customer (yes, often in big letters on screens behind the speaker), ViPR just moves lock-in somewhere else. While the customer may be henceforth free to buy storage hardware from any vendor or white box supplier, the software layer that is ViPR, knitting it all together, will not be something they can reverse out of with ease.

Tony Lock, product director with Freeform Dynamics, concurs. “The management layer is the stickiest of all. It’s difficult to remove if customers build processes into it, and especially if several management tools are layered together.”

EMC more or less says as much too. In an interview with the brains behind ViPR, president of advanced software Amitabh Srivastava, I asked the question: “A disk vendor allowing customers to use any vendor’s arrays? What’s in it for EMC?”

“It allows EMC arrays into new markets, such as big data analytics, Hadoop etc. And it adds new capabilities, overlaying new features onto storage arrays via software,” said Srivastava.

In other words, it’s a new sticky layer of product that will lock the customer in. Srivastava couldn’t tell me what the charging structure would be for ViPR, but one would imagine it’d need to make up for decreasing disk revenues elsewhere.

So, we know what it’s for and how it furthers the interests of EMC. But will it succeed? That, of course, remains to be seen. For it to fulfil the promise of tying together multi-vendor systems it needs adapters that will facilitate those connections. So far, there is only the promise of adapters for some NetApp arrays.

Nearly everyone seems to think ViPR is a good idea. The question is, can EMC convince the rest of its storage competitors to play ball and let it gain revenues from its management layer while their storage arrays sit below it?

Freeform Dynamics’ Tony Lock says: “The big problem will be getting third parties involved. That could make or break it. It’ll take political will and diplomacy.”

Only time will tell. It could be that ViPR emulates a standard to which no-one subscribes, SMI-S springs to mind. Or it could be that EMC’s 800lb gorilla stature means it can muscle its way in via its market leadership position, a la Microsoft.

Or there could be a third way; ViPR gets used for some things and not others. At EMC World the company had a potential customer on hand, Columbia Sportswear,  to talk about its intended purchase of ViPR but it only really planned to use it as a quite clever storage virtualisation platform.

With 4,500 employees in 28 countries and four datacentres; two in the US, one in France and one in Hong Kong with a mixed VMAX, VNX and Isilon environment running to about 1.5PB, Columbia doesn’t have a cloud or big data use case in mind. It does, however, see the possibilities of uniform storage settings in the ViPR layer across different array environments.

Mike Leeper, director of global technical infrastructure with Oregon-based Columbia Sportswear, said: “We have a 98% virtualised server environment and we see the benefits of virtualising the compute layer. But we’ve looked at where we’re being held back and storage is one of those areas, so we’re interested in how we could make workloads portable and not tied to specific storage hardware.”

ViPR appears to be a completely new product category, merging storage virtualisation with the cloud and big data. But, only time will tell to what extend ViPR will succeed, or to what extent the rest of the industry will emulate it so that we end up with many ViPRs. Storage companies are notorious for building walls around their products but perhaps EMC hopes it can be the capstone that sits atop a new type of data environment. It’s going to be an interesting development to watch.





Read more on Virtualisation and storage