HDS customers include more than 70% of the Fortune 100 and more than 80% of the Fortune Global 100. It claims to be the fastest growing storage vendor in the industry and that it grew twice the rate of EMC in the last quarters.
I’m led to believe that Hitachi Data Systems is considered a model for how Hitachi itself would like to operate. What do you put that down to?
We’re producing results. In the last five years we’ve seen some really good results, for example we had excellent growth even in 2009 when the rest of the competition and even Hitachi itself struggled. I think it’s about vision, and at HDS we paint a five to ten year picture of where we want to focus, how we want to align our strategies. We have a positive culture and believe in building a sustainable culture for the long term. I think we’ve got a good balance.
You sell a considerable amount through channel partners, right?
We do, and I think our channel partners value the fact that we are clear about our strategy. We expect channel partners to put resources into the relationship just like we will. Our indirect revenue actually surpassed our direct revenue for the first time this year. So I think our partners are comfortable with the vision.
HDS has made a number of acquisitions: BlueArc, Archivas, Shoden. Have you been pleased by the integration of those buys?
Absolutely, I think the integration of all of those went very well. BlueArc had been a partner for five years before we bought it. BlueArc is really in the unstructured data area, where we have been seeing this explosion really, and we felt their technology could really complement what we were doing in structured data.
Archivas was also in the unstructured space but more at the block and file level, so with BlueArc we think we have the right pieces for what we call the content cloud. Shoden really gave us some market share, and they had been a partner for ten years also.
Are further acquisitions likely or do you feel you have the key pieces?
When we look at acquisitions we ask how they will align and what they bring to the table. I’d say we have constantly got our eyes open for potential acquisitions.
Is storage virtualisation as hot for HDS as server virtualisation seems to be for firms like VMware, Microsoft?
We’ve been there and done that, to be honest. We have been in [storage] virtualisation longer than anybody, right from the mainframe space. So our storage is not only made for virtualisation but it has been truly tested in the most demanding environments.
So yes, we’re seeing lots of traction there. Prices generally are not coming down, and in many firms the supply [of capacity] is there, so virtualisation can help to drive up utilization rates. Our virtualisation offering is pretty powerful; Gartner said we have the best virtualisation on the planet, offering 25- to 60% utilization improvements. It means people don’t need to buy storage for two or three years, and we virtualise us, EMC, NetApp and so on.
But we also believe that the real value comes from turning data into information: that’s critical. Can you manage all your data, find all your data, create value from all of your data? How does your organisation turn data into information?
What does the term ‘big data’ mean to HDS?
Big data means a huge opportunity for HDS. There’s a lot of hype around big data and there’s some suggestion that it’s a repurposing of analytics that have been around for 30 years. But while some people are skeptical I do believe that technologies like Hadoop are being adopted because people need new ways to analyse data. The growth of data, particularly around unstructured data, requires new tools.
For instance companies like Walmart want to know who’s blogging about them, Tweeting about them. Think of the amount of data created in industries like health and life sciences: who’s going to analyse it all? You need a whole new governance model for unstructured data – music, video, social feeds, machine data – being created by thousands or even millions of transactions outside of the data centre. The scale is just huge.
We can now index all data no matter what application created it – as we store it we index it. So then you can search all of the data independently of the application, and since we virtualise the data too you can now discover it in the cloud.
What role do you see flash memory playing in the storage market? There are a number of flash-based firms like Whiptail, Kove, Texas Memory Systems coming onto the scene.
You will always have different levels of media. You call it flash, I call it cache. You need to embrace them all, and we do. Non-spinning disk is very efficient. A lot of transactional data should be on flash or memory, then if it’s not accessed very often it should be on disk.
Data has a viscosity and it’s always moving; depending on the type of data you may put 95% on non-spinning disk, or in the cloud, or it may be only 5%. There’s been a lot of hype around SSD, we have said consistently that we’ll put the data where it needs to be. For example we can store data in an Oracle [database] row on flash, on a row by row basis: we can optimise it for the business need.