Hitachi Data Systems CTO Hu Yoshida said he sees the need to "transform the business" to compete in the disrupted...
and slumping data storage infrastructure market.
"If the infrastructure business is not doing well, doubling down on it by combining with another infrastructure company is not really addressing the problem," he said during a recent interview covering HDS storage and the overall storage industry.
What's your strategy to compete in the disrupted storage world?
Yoshida: The infrastructure market is not going to be a growth market. Instead of trying to compete on infrastructure, we're going to have to compete on application enablement. A large online company is the largest purchaser of our flash drives. What we did there is displace, or reduce, their infrastructure spend by 80% by going to our HCP content cloud. So, we are looking at application enablement and analytics enablement. That was the reason for the acquisition of Pentaho, so we're not competing on infrastructure. Infrastructure is not a market to be putting our eggs into.
Now, infrastructure is still important, because the right infrastructure -- like our flash modules -- enables the rest of the stack. But it shouldn't be the prime concern of IT. And, in fact, we can even use the Amazon infrastructure as a lower tier for our content cloud, because we can interface to the content cloud through [Simple Storage Service] and open interface APIs.
That's the strategy we have. It's not just to try to make a bigger infrastructure company, like Dell and EMC, or to get rid of other pieces of the business, like Hewlett Packard Enterprise. We have to transform the business.
How do you view the impact of public cloud on the overall storage market?
Yoshida: [Amazon Web Services] AWS does not buy storage from the infrastructure vendors. They go to [original design manufacturers], and they have their own disks made for them. Analysts say that for every dollar that goes into AWS infrastructure, it displaces $3 to $4 of infrastructure spend. That's a huge amount that's being taken out of the infrastructure market. I think Amazon's expected revenue this year is $8.1 billion. So, if you consider displacing $3 of infrastructure spend, that's $24 billion of addressable market that IBM, EMC, NetApp [and] HDS cannot address, because it's gone into the cloud.
What do you think are the right ways and the wrong ways to deal with the changing storage environment?
Yoshida: It's not just about the storage anymore. You have to move up the stack. We are hiring different types of people now. We're looking at analytics, application development, containers, Docker. We're looking more toward the application enablement side of the business. Yes, we still have a storage business, and we still have engineers working on that -- developing the next flash drive, developing enhancements to our storage controller. More of it's focusing around software-defined and automation. It has to be a converged solution, not the one-off storage device that we're going to sell on its own. And it has to be part of a solution.
What's your take on software-defined storage?
Yoshida: What I see as software-defined storage is about communication between the infrastructure and the storage system and the application. It is a communication, but it's not going to be any better than the hardware that's underneath it. For instance, VVOLs [Virtual Volumes] from VMware enables me to publish my unique capabilities. If I do active-active, virtualization, I could publish that up through the VASA interface so that vSphere can see my capabilities and define a virtual volume utilizing my unique capabilities. From my standpoint, it also enables me to actually see the virtual machines and provide resources for virtual machines, rather than just publishing RAID groups up to the server and letting them carve it up.
Do you think storage becomes a commodity in the software-defined world?
Yoshida: No, I don't think it'll commoditize storage. It enables us to publish our unique capabilities and differentiate ourselves so that the upper layers can use our capabilities much simpler. For instance, the intelligence within my flash drives gives me better durability, better security [and] the ability to scrub devices. But I still interface up to the standard interface.
Is there a place for commodity server-based storage?
Yoshida: There certainly is a place for that -- the hyper-converged, the scale-out storage compute nodes, where you have storage and compute, and the software is handling the global file system. You don't need to have RAID. You just fail over. There's a place for that, but it's for a MapReduce or a scale-out type of application. The data lake is a great place for these scale-out commodity storage compute nodes.
If we consider HDS storage today versus next year at this time, what will be the main difference?
Yoshida: We have an overall corporate strategy with Hitachi, called Social Innovation, where we are moving toward the Internet of Things (IoT), trying to build smart cities and provide more insights into data centers, telco [and] automotive. Because our parent company has a lot of vertical expertise, we want to be able to leverage that and provide solutions around social innovations. We have something called Live Insight for Telecom, so we can monitor and optimize all the traffic. We also have that for data centers, because data centers are an IoT environment. Every application, every piece of hardware has a log. If we can collect all those logs and start to do some correlation and analytics on that, we can probably enhance the availability of the system. We can reduce the power [and] cooling. You may have data centers in Singapore, Hong Kong and London, but we load all that data into the cloud. And through a portal, you'd have one view of all those data centers. And then, we could run analytics in that.
What trends do you foresee with IT users?
Yoshida: IT is going through a transformation. Every CIO is talking about transformation and how they're going to have to change. The world of storage administrators, server administrators, network administrators is going away. They want somebody with higher-level skills -- cloud enablement, analytics [and] things like that. And the major task is: How do you transform your people? Some people are going to be the hardest thing to change.
How long will it take for the storage administrator to fade away?
Yoshida: I think you're going to start to see that right away. A couple of weeks ago, I had a customer in, an electric company, and I expected to see the CIO. He didn't show up. In the course of the conversation, the IT director told me that they were going to do an Oracle-to-SAP conversion, and they hired an analyst firm to help. In the midst of the study, the analyst firm said the CIO was not adding any value. Within two weeks, he was gone.
I asked the rep to see if this is a trend that he saw happening in his area. He said that two other companies did that. One was city administration. Another was a retailer.
Maybe something else was happening, because two weeks is a short time to get rid of a CIO, but that is the danger, if you're not focusing on transformation.
On the other hand, there are CIOs that are being given more responsibility. Now, we're talking about chief data officers. The data officers are looking at analytics and the applications around that. So, there seems to be two trends. One is where the CIOs are really cost centers. They seem to be going away. And where there are really innovators, they're getting elevated into board positions.
HDS' Yoshida discusses virtual data storage challenges
HDS storage launches new all-flash array
Lowdown on HDS VSP G1000