IDC's Richard Villars knows a thing or two about storage. He has spent 19 years with the analyst firm assessing the emergence of network technologies and general business practises. Now vice-president of IDC's information infrastructure team, he advises on the impact of tiered storage, blade servers and regulatory compliance on an organisation's information management.
According to Villars, the single biggest technology trigger affecting the storage landscape today is virtualisation. While take-up is initially being driven by physical storage consolidation, other key concerns increasingly influencing customers adoption include backup consolidation and eliminating de-duplication, cost effectiveness, improved performance and reliability and disaster recovery.
This shift in market dynamics will see the advent of more role based, applications-centric storage and force vendors to focus on specialised services and collaborative software, he claims. Villars recently shared his insights with a selection of Hitachi Data Systems' local partners. He spoke with Nadia Cameron afterwards on the changing face of storage and what resellers need to focus on.
Q: What is the most important technology affecting the storage industry today?
Richard Villars (RV): Virtualisation is affecting the whole IT industry. In some ways it's unique: if you think about PCs, the Internet, or LANs, they were all things that business units took and did their own things with. IT then got dumped with it and had to rationalise. Virtualisation is completely the opposite.
Enthusiasm and interest is being driven by the IT department. In some ways, they're taking this up because it makes their lives easier. They're not getting business units walking in and saying 'we need VMware or storage virtualisation'. This means it's happening faster and is more invisible to the outside world.
Q: Who are the biggest influencers in the storage market?
RV: From the outside world, it's VMware. In the broader sense, Microsoft and some of the major application players still have a lot of influence. And increasingly today, what you can build and deliver is dependent on component people such as Intel, AMD, and the disk drive manufacturers.
But it's no longer about a manufacturer placing orders with small suppliers. It's about having a co-operative ecosystem. For most of those classic storage suppliers - EMC, HP, Hitachi, NetApp, Sun, IBM - increasing their value and strength is around delivering a storage solution to the customer. That could include multiple tiers of storage as well as multiple functions such as disaster recovery, performance, or fixed content capabilities.
Q: Is that what's driving the many acquisitions occurring in this space?
RV: Yes, that's why you're seeing the acquisitions that you are seeing. For example, Hitachi's decision to purchase Archivas was based on its very strong software technology position. For Hitachi, it's about how to create role-based storage specifically for that archiving space. It's similar if you look at the buys EMC has made across the board.
One of the challenges right now is that because some of these technologies are so far ahead of the curve, it's hard to tell which of those buys is going to work and which may not. But it's less about buying hardware. The exception is companies buying storage hardware. I do think we will see interest pick-up in some of the file virtualisation parts. Because these are actually switches, there's a network element to it which does raise the question of who buys them: do they get bought by EMC/Hitachi, or do they get bought by someone like Cisco? So now we're seeing the networking side having a big influence on storage as well. And increasingly, the capabilities of the network are defining what you can do with your storage.
Q: What are the top selling points resellers should be using to pitch the storage argument?
RV: For the next year or two, it'll still be about consolidating servers and storage. But I do think, especially these days, that more of the pitch is about availability and disaster recovery of data. Availability is the key thing: waiting a day for the tape to come back and recover an application is unacceptable, even in a small business. And it's not just transactional data but also the collaboration. This was one thing businesses discovered in Florida after the hurricanes hit and their data centres were knocked out. Employees were sitting at home, with laptops and broadband access, waiting to do work, but they couldn't get back into the data centre to recover the exchange and the basic business apps. The employees couldn't do their jobs because those non-mission critical applications couldn't be recovered. I think the change in mindset is recognising that disaster recovery means the collaboration tools, not just the business systems. And I think we as a storage industry need to emphasise that point.
Q: How does that tie into the skills resellers need to develop and brush up on?
RV: The skills required are becoming more processes-based. Resellers need skills not only in break/fix, but also in what steps are needed to get the application up and running. They need to know to reconfigure the network, reload the data, and understand the procedure in order to get the application up again. Resellers are a big part of building that checklist and in a lot of cases, following through on it.
Q: So in five years' time, we're not going to be talking about storage but instead about particular applications such as disaster recovery?
RV: Disaster recovery is one of the main applications for storage. Five years' ago, if you said disaster recovery, all you thought about was people like Telstra and the top five banks. But now it's applicable to small regional companies. This is partly because the definition of what a disaster is now is so broad.
I think the other thing you will see in five years' time is a massive increase in data mining. It may not be for small businesses, but even mid-sized organisations are going to want to respond to customers by mining internal data. Making copies of that data so you can mine it without disrupting ongoing operations is a big storage concern.
Another thing you'll see is application awareness of storage. The problem we've had with storage in the past is that it was very utilities functional. Replication, for example, was a utility that administrators used to do a migration or a fix. Now storage is almost like an operating system - and it is about services. I might have a replication service, for example. An application can take advantage of this to either ensure better productivity or to ensure data is in the right location at the right time. To do that, that service has to stop being a utility: it has to be an integrated function with an exposed API that people can use. Virtualisation is the first step in making that a reality.