Tuesday Nov 26, 2024
Tuesday Nov 26, 2024
Monday, 8 July 2013 00:00 - - {{hitsCtrl.values.hits}}
Hitachi Data Systems’ Vice President and Chief Technology Officer Hu Yoshida is at the forefront of the company’s success and currently leads the company’s effort to help customers address their Data Life Cycle requirements to address compliance, governance and operational risk issues. He was instrumental in evangelising Hitachi’s unique approach to storage virtualisation which leverages existing storage services within the Hitachi Universal Storage Platform and extends it to externally attached, heterogeneous storage systems. Yoshida is an icon within the data storage industry and his blog was recently ranked among the “top 10 most influential” within the storage industry by Network World. In 2006, he was named “CTO of the Year” by Jon Toigo and in October of 2006, Byte and Switch named him one of “Storage Networking’s Heaviest Hitters”.
Yoshida was in Sri Lanka recently in support of the company’s local endeavours, where the Daily FT sat down with him amidst his busy schedule to talk about the company’s systems and what they have to offer Sri Lankan businesses.
Following are excerpts:
Q: What brings you to Sri Lanka?
A: Well it’s really to understand the business here, the customers and our partners and try and raise visibility of our company here and in the region. We’ve had quite a bit of success having been here only since 2011, especially in the banking areas and telcos; we’ve seen a lot of opportunity and growth here.
Q: You have been responsible for defining the technical direction of Hitachi Data Systems and also helped define HDS’ main strategies for the company’s technology offerings. What do these strategies entail and what in your opinion is the future of data storage and management?
A: What we’re facing is an explosion of information and data, and more and more new types of data for instance the smart phone. That has probably brought another billion people online in the past few years and they’ve become consumers and producers of information and now we’re looking at machine to machine types of information with smart grids and stuff like that and so we have to be prepared to handle much bigger volumes of data and be able to handle it much faster in terms of being able to search and find things within all this data and then turn that into useful information. So we’ve had a basic strategy for some years and it’s based on virtualisation. To be able to grow and expand quickly, you’ve got to separate the infrastructure from the data so you can refresh, migrate and change the whole structure without impacting the access to the data or the applications. The next step is to separate the data from the application because applications also age and need to be refreshed and that’s often more disruptive than changing infrastructure and to do that, when you separate the applications from the data, it’s just a bunch of bits and you have to put that into a container with metadata that describes it and policies that govern it. Once you do that, then you have an object, so the emphasis will be more and more on object stores where we have metadata that describes the content because independent applications generally can’t be used by other applications.
The last step then is trying to find the information in all these different data silos and what we have to do is be able to see the intersections of that to be able to search. So it’s kind of a three step process virtualisation of the infrastructure, virtualisation of the content and then virtualisation of the information.
We’ve developed a portfolio of products that helps us to do that.
Q: HDS has been described as one of the most complete and feature-rich packages available for high-end enterprise storage. Tell me in your words what is special about the platform and how important is it for companies to engage efficient storage virtualisation systems?
A: The basic product is our enterprise virtualisation platform. This is not an appliance, this is a virtualisation control that has all the availability, the scalability that you need in an enterprise because if you think about what’s happening with virtual servers, these X86 virtual servers now, they’re just like mainframes because you’re stacking up all these applications so you need something that you will be able to scale up and also its high availability.
The architecture we have is unique, nobody else has this and it is more of a grid architecture. It has an internal switch and I can add to that switch processors that do the control, I can add ASICs that do the data movement and I can add cache all connected through the internal switch so that I can keep adding processors if I need more processing power and more data movers or more cache and it will scale up. If you look at what everybody else does, they provide you with a processor that handles the data movement as well as the cache with one node, and if the application needs more processing power, it can’t get it out of that node. What they will do is cluster it with other nodes, essentially called a scale out but that doesn’t help if you unless you redistribute the workload and we don’t have to redistribute the workload because it’s in a great grid. What they do is they put these nodes on an external switch and so everybody else does a cluster of these nodes even IBM and EMC. The thing about the virtualisation we do is that as we attached other systems to it, we can enhance those systems with all the functionality of our enterprise control. It could be a smaller, low end product or legacy product, but once it is attached, it has all the enterprise functions. On the other side too we can attach filers, we can attach content platforms and BTLS which then provides a common platform for the applications as well as the data beneath it. This allows us then to manage all this with one set of tools. We can manage our servers, our files and our content all with one set of tools and when you’re talking about this explosion of data into Exabytes, you can’t match them in size. You’ve got to manage them as one pool of resources. So those are the advantages of our architecture and virtualisation.
Q: What in your opinion are the benefits of heterogeneous vs. homogeneous storage and how have you managed to incorporate it in your systems?
A: All storage do not have the same requirements, or all data do not have the same requirements. Not everyone needs Tier 1 availability and performance and much of this data becomes inactive or doesn’t get refreshed much so you could tier it down to lower cost tiers of storage. So that’s why virtualisation which enables you tier storage and automatically move data across these tiers of storage. That allows you to have heterogeneous systems. We provide low-cost storage as well but we can also use other people’s storage. In addition, as IT systems grow, there are a lot of legacy systems that people need to incorporate into this pool and this allows us to do that. The way we do that is through the virtualisation, so this allows us to add in different types of storage. You know, like in the mainframe, that used to be a logical stride, we had mainframe storage and now what we are able to do is to put our enterprise system there that talks the mainframe language but beneath it we can put heterogeneous systems. So you get better cost, better availability and you get to lower operating costs, so many benefits
Q: With Cloud computing being a catchphrase all over the world, where does it fit into the HDS equation?
A: Cloud is where I believe, everybody will have to go eventually, because what Cloud does is it gives you on-demand infrastructure. Today, the way people buy things, they project out what they need for the next five years and they buy what they need for the five years at today’s price. So they buy all this capacity that they paid for but they’re not really using until they get to the end. Then of course when they need to go on to the next technology they have to buy that to replace this with the new but they will always have this overlap because they have to move the data. By the time they finish this migration, they have to start on the next. That’s a very inefficient way to manage your infrastructure. So by going to a Cloud, you can buy what you need when you need it at the current price and you don’t need to project out what you need for the next five years; it’s pretty hard to do. It’s also very difficult to do because most application users don’t know what they need; also they don’t tell the IT people the copies they need to make. Generally they underestimate and they’re all scrambling around in the end to pick the best. So not everybody of course wants to use the Cloud because they’re concerned about security and things like that but there is Private Cloud and to do this you need three basic technologies. One, you need to be able to dynamically provision the storage, in other words, put more storage in there and the system dynamically format it into various pages. Secondly, you have to do a non-disruptive migration where you don’t take this overlap and we can do that with our virtualisation system. The third thing is again the virtualisation, so if you’ve bought something in the fourth year of its five-year depreciation, you can continue to depreciate it because the functions are at the front. You don’t lose the depreciation of the asset, so that helps you leverage your assets. So those types of capabilities are there so that you can have this dynamic on-demand growth. There’s also the financial piece to think about, because somebody’s still carrying the risk of the inventory. So yeah we could manage that through creative leases, maybe serviced or private Cloud. So Cloud is definitely in the future and that will eliminate all this overlap and waste.
Q: Innovation is an essential survival skill for global businesses. How does your company approach innovation and staying ahead of competition?
A: Innovation is, I would say within our DNA. Hitachi is a company that is over a hundred years old. When it was established, three principles were set; one was harmony, the second was sincerity and the third one is what we call “pioneering spirit” which today means R&D and innovation. That has sustained us for a hundred years through wars and depressions and all that and guided us. So innovation is one of the key tenets of our system. Hitachi is an engineering development company, we don’t buy off the shelf processors and try to fit them, we build ASICs that does the data movement and our filers are the same.
In Hitachi Data Systems in Santa Clara, where the headquarters is, we established an R&D site there in the year 2000 and it brought in researchers from different disciplines all over Hitachi and since then hey have developed over 380 patents which is very productive. However last year we changed that from being an R&D lab to an innovation lab. The difference is that the measurements is no longer on the number of patents we file, its measurement is now on failing fast and trying things. Sometimes we learn more from things that fail than things don’t and patents that just sit on the shelf. So we’re driving that at a much faster pace. All the research labs at Hitachi in the last years are also working much closer together; it’s something we call Kenkai it’s a Japanese term for something we do where all the research teams come together on a regular basis and share ideas. Where this is really helping us is in big data, what we’re looking at is the big data of tomorrow, not just business data with some social media around it but more of a machine to machine focus like train systems talking to ticketing systems and so on. To do that type of big data you need deep expertise in transportation, energy and sciences. Fortunately our parent company has that capability. We provide the information infrastructure for that. So like one of our recent projects where we provided the hybrid trains for the London Olympics. There is a lot of big data analytics going on managing these trains, the maintenance and also the ticketing and what that took was Hitachi divisions that did train systems and control systems plus Hitachi Data Systems that provided the data storage infrastructure for that. It’s given us a new accelerated level of innovation because now we’re looking outside of infrastructure and data systems and focusing on solutions as well.
Q: You joined HDS in 1997 after a 25-year stint at IBM and having previously served as a US Marine as well. How have those experiences helped you in developing your style of leadership?
A: In the Marine Corps, where I was an officer, we were trained to first take care of their men, look after their welfare and then secondly to lead from the front. When I joined IBM it was a very similar culture at that time because the key thing was respect for the individual, by that way you get the most out of the individual. Also I joined IBM at a very amazing time where there were innovations coming out one after another and I also had the good fortune to work with very good engineers who were very creative people. That gave me a good background in storage systems and also in storage software and then it also gave me the opportunity to work in Japan for IBM where my job then was to attach IBM storage to Hitachi and Fujitsu operating systems. So then I got familiar with Hitachi and as I looked at its architecture and I was very interested in it and that why I joined Hitachi.
Q: The global financial crisis has affected the bottom lines of companies across the globe with Europe especially feeling the crunch. What is your company’s secret for dealing with and weathering the storm?
A: With virtualisation we create great efficiencies. Most of the places where we come in and virtualise the systems and we save about 40% of the capital and operational cost, so during downtimes is when we become most successful and the customers come to us. You know we had the downturn during the dotcom bust back in the late 90s. During that time, that’s when we had lightning product and that’s when we started to gain market share and that’s also the time we exited the mainframe business, we gained greater market share because with that system we could replace six to eight competitive frames. Now we can connect more servers to storage frames and then we move on to virtualisation so we’ve created more efficiencies especially during downturns when people have limited or no budgets. That’s when we become most successful l. Last year, year on year we grew 5% and it wasn’t a good year for storage on a whole. We are ISO 14000 certified, that means at the design phase every part is designed to be more energy efficient. Our supply chain vendors are also all certified, our factories reduce the carbon footprint and we also manage the distribution and maintenance. If you compare our product to our competitor’s product, we save maybe 20% to 30% or more. The recent product we just announced is the HUS DM it is a modular package but has an enterprise capability. The controller is only eight inches by nineteen inches and to have the equivalent capability, our competitors have to have a full 45 U frame plus with the EMCs case they have to have another frame for the discs because in these controllers you know we have cache and that cache is volatile. So to protect that cache they have to be able to download it quickly in the event on a power failure and EMC downloads it to flash drives through fibre channel drives. They require five fibre channel drives per path, 16 back end paths that makes it 80 drives plus eight more for spares, that makes 88 drives they have to have just to protect their cache. IBM takes a different approach, they lock everything in caches with batteries, so when you open the frames, it’s full of batteries. They also need some discs for booting and so forth. With us we don’t require any of that because we download our volatile cache to non-volatile flash so it’s a very small package. So that way our not only does our controller use much less space but it’s 0.6 kba versus a full 6 kba for IBM and EMC. So that’s how we save on top of all the other virtualisation things but also saving the cost of energy and floor space which is getting to be more of a cost factor.
Q: How much potential does Sri Lanka have as a market for your systems?
A: As the Sri Lankan economy grows we see a lot of enterprises looking for newer technologies to help them expand their businesses and topics like disaster recovery and consolidation become more and more important. We’re very happy that we have been able to do very well in telecoms and banking and those are the many opportunities that we see here. Of course there are many more sectors that are also booming for example in manufacturing and government and perhaps also in the tourism sector.
I think this is sort of a green field for us. In other countries, you see a lot of legacy equipment that people don’t want to change but here in Sri Lanka people are only going for new technology. Within the last three years we’ve sold a lot of our high end systems and normally people would think that this is Sri Lanka, and as a small country that they would want the mid range product but customers here are going immediately to these very high end systems, virtualisation with disaster recovery replication, all the latest technology is here in Sri Lanka. It’s not a legacy country where everybody is holding on to the old technology. In fact, your largest commercial bank uses systems that are equivalent to the systems used by any larger international banks such as Citibank etc. Banks and telcos need to be competitive on a global scale so they need that technology.
Q: What are the biggest challenges the industry faces right now and how would you address them?
A: Well certainly, everybody says there’s a growth explosion of data, that’s one of the challenges and the other thing is the budgets where they want to do more with less. The other thing is the agility that is required today; the world is changing very quickly. You see on the stock markets the flash crashes where the market drops 400 points. How do you react to that? Agility is a big requirement. So again our basic strategy revolves around virtualisation and to make sure you are not tied to infrastructure and not tied to applications so you can move and scale and do all this with wider availability and these are the things we need to address. So I think we are in a sweet spot because of our leadership in virtualisation and our ability to manage our performance and tools. I would also like to thank all of our local customers that we have had the pleasure of working with here in Sri Lanka they have shown a good understanding of the our products and how it can help them do things better. Sri Lanka has a great future ahead of it and the one thing I’ve noticed here is that the streets are extremely clean and well maintained and that in my opinion is a very good sign of a people who take pride in their country.
