As the PC market stagnated over the past three years, tech giant Dell has been reinventing itself as a software, storage and services company, along with its acknowledged prowess in hardware. To do this, it has been not only on an acquisition spree but has also been investing heavily in 'practical' research and development. The man leading the company's R&D efforts is Jai Menon, an IBM veteran. Menon, who was in India recently, spoke with Venkatesha Babu on 'practical research' and India's role in it. Dell has annual revenues close to $60 billion and employs about 27,000 people in India, a quarter of its global workforce. Edited excerpts:
Q- Dell has always been seen as a fast follower and not a true innovator. What has changed?
A- I joined Dell about two years ago. The first nine months I actually spent as CTO of the enterprise solutions group and initially had the same impression of us being a fast follower and so on. It took a little digging around inside Dell for me to figure out that actually we have a very long history of innovation. But, you know, maybe some of it could have been better brought out. I'll give two or three examples.
The move from using proprietary fibre channel type networks for storage to things like iSCSI (Internet Small Computer System Interface, an IP-based storage). Dell made the right moves at the right time and got into that business and now it's number one in that space. That's a really important aspect of innovation, which is sort of guessing where the puck is going to go and making the investment at the right time.
Q- But didn't that happen because of an acquisition?
A- Absolutely. By the way there is nothing wrong with that. You have to innovate in different ways. Some of it is through acquisition, some of it is through organic, you know in some ways, it sort of doesn't matter, because there are smart people everywhere in the world and if something has already happened and it's already been built, there is no reason for you to go build it again. Right?
So that's one, and, of course, once you acquire the company, then it becomes organic innovation that follows. The fact that we are still number one is because of continuing innovation in that business. The Dell Data Center Services (server) business is another example which was purely organic. That was not an acquisition, but that's another example of sort of guessing that hyperscale clouds and sort of skating to where the puck is going, driving, creating that business actually before a lot of other people got into that business and, of course, once again in a hyperscale cloud business we have later.
It used to be 53 per cent share when I was CTO. I don't know if it's exactly that same number, but it's in that ballpark, which is still a large vendor in that space. And actually time and again I found some of these things, some of the flash standards, I found that Dell actually had been one of the key founding members of some of these groups and some of the early flash products actually had really strong performance. So, you know, I found example after example of that kind of innovation.
So I feel that there has been a long history of innovation at Dell, not all of it well publicized and well understood externally.
Q- Unlike say a Microsoft or an IBM which believe in investing in fundamental R&D, Dell's emphasis seems to be on a practical, more of a result-oriented, more about business-generating kind. Your comments.
A- My point of view is that at both those companies, too, actually predominantly, it's about practical innovation, it's about innovation that at the end of the day makes impact on the business. There is a different filter applied to what is described externally, but if you actually look at the overall R&D budget and ask yourself how much of that is actually in some of these truly fundamental kinds of things, I mean, I joined IBM at a time when that arrow was shifting from one place to the other. So I think that, if you were to go ask a lot of people today, I would say that, I think they do a lot of practical innovation, which is a positive thing to say about…
My philosophy has been to not let me go work on things that may or may not see the light of day 20 years from today. Fundamentally, my point of view is that you want to tie things back; it may take some years, that's okay. But if you can't.
Michael (Dell) provides absolute leeway. So, for example, as part of the research organisation, which is the thing that I now lead, my mandate is (for projects which could yield results in) 18 months to five years. I'm trying to make sure that there is lot of innovation going on in each business unit. Each business unit is thinking about the next generation or the next to next generation. Our work begins where some of the R&D within each business unit ends, so that if they are already thinking about what the next product looks like, I'm going to start my thinking beyond where that activity ends.
So if you look at my organisation, which is the Dell Research Organization, of course, I have a broader role for the overall R&D as well, but if I were to just focus on the charter of my organisation, I would say I have three things that I'm trying to do. Do organic research that is long range, it's pan-Dell and is disruptive.
Q- Could you give some examples of such disruptive thinking.
A- It could be a new form-factor. If you are focussing on what the next tablet is going to look like, you might miss the fact that wearable (computers) are coming along. A lot of people missed the importance of flash. I mean, flash came along very suddenly. Now we are tracking technologies that come after flash and they are going to be important in data centres by, let's say, 2018.
So, they could change dramatically what is possible with servers. You have to rethink how these kinds of memory and storage devices get attached to servers, those things could be quite disruptive. Let's say Flipkart on Diwali day, has five million, 10 million customers doing transactions. How can we, in real time, be able to detect whether a fraudulent transaction is taking place or not?
And you may not have the time to go to Flash and access the data. Next generation technology is going to be 50 to 1,000 times faster than flash. How do we leverage these technologies? What kind of applications might be possible? So every year or so we release Dell's technology outlook. It's a Dell point of view, but it's informed by our interactions with other people inside Dell, of course, but also people outside Dell, people from university community, people from the analyst community, people from other research organisations, customers.
Q- Another example of such technology forecast and how it will play out?
A- A good example is, in research we actually have been looking at the telecom industry. So whether it's AT&T and Sprint in the US or whether it's Reliance or Bharti Airtel or something like that here, if you look at the current networks, what you'll find is that they build it up using specialised hardware equipment. It's typically proprietary, you can only get it from that single vendor, all the boxes that make up that network that they have, the mobile phone network for example. All of the boxes, there is, 10 or 15 different types of hardware devices that they have to buy in order to build up that network that supports your mobile phones and also enterprise connectivity etc.
What is happening in that space that Dell research was influential in bringing this up within Dell was instead of buying these specialised pieces of hardware, we have found that you can actually create the same function as software running on standard servers.
Q- Just like it's happening in the data centre space?
A- It's just like that. The data centre journey started 10 years ago, the telecom journey is starting now. Today they all buy specialised hardware. The reason is that the kind of functions, the kind of applications that are running on those special purpose hardware is different than the data centre applications. And it requires what -- and, in fact, we call it a high-velocity cloud, because it requires you to push a lot more packets per second using standard servers and then a standard server might typically be able to do.
So, for example, if you take a standard server, maybe you can push 200,000 packets per second, but what these things might require to push millions of packets per second. Part of the research that we did was to create some specialised layers of software within the server that allowed these applications to get the kind of performance they need. So, five years ago, three years ago, would not have been possible and that's why they haven't done it yet.
Q- Which are the companies likely to be impacted in this shift?
A- I think Huawei is an example of a company that sells into that space. Cisco does some of it as well. Now, all these companies also want to move to this new model. But because they have an entrenched business already, the rate and pace at which they will want to move is a little bit different than the rate and space at which a company like Dell will want to move.
We're already strong in servers; we sell a lot of servers. And we also sell, now with our Dell software, we also have cloud software and so on. In fact, the market is really coming to where we are going. That's an example of big industries like $160 billion that all buy special hardware today. So imagine it's a ten-year journey, from where they are to where eventually it's like the data centre. Eventually, 80 per cent will get virtualised, you still have 20 per cent still buying special hardware, but 80 per cent will get virtualised. But it will probably be a 10-year journey starting next year.
So, it's sort of in the POC (proof of concept) phase now. I expect over the next year there will be little roll outs in little rural areas somewhere to iron out kinks and then roll out everywhere.
That's the power of what is possible. It is still going to be a journey over eight or 10 years. But imagine that only 30 per cent of this $160 billion market moves over, that's still what, $48 billion or something. It's still a lot bigger than the entire server business, which is $50 billion today. Just imagine if I can move 30 per cent of this over that's doubled the server business.
I keep telling my team we should either make a product better, make a new product or make a new business. It's easier to make a product better, it's a little bit harder to create a new product, it's even harder to create a new business.
A lot of disruption is happening in cloud, big data, new kinds of flash memory. In many ways the thing I like about Dell is, it reaches out and really tries to democratize technology.
I mean, other companies might create technology but it's only accessible to the biggest businesses, it's expensive, it's complicated. And a lot of what Dell does is, it takes that and makes it more accessible, democratizes the technology, makes it easier to use, makes it more purchasable, if you will.
Q- To that extent the DNA of the organisation hasn't changed from its inception?
A- No, but it takes a lot of innovation. I mean, there is different kinds of innovation. It takes certain kind of innovation to build the biggest, craziest, baddest thing that only three people in the world can buy, that's one kind of innovation. But I'm so excited and charged by a different kind of innovation, which is to say, how do I take this thing that only three people can use today and how can I make it usable by a million users. It's a different kind of innovation, but it's just as hard, just as complicated.
Q- From your perch, what are the key technology trends you can call out for us?
A- I'll give three or four of these. One is this -- this telecom example is the broader claim that I like to make is that, we believe in what we call software-based data centres. A lot of people talk about software defined data centres, I'm going beyond that. Software based is a Dell terminology and essentially it's saying that our servers get faster and faster, a lot of its software on these standard servers is going to replace. So things like storage, networking, firewalls, all of these things eventually move, migrating to servers and then software or servers. And that's what we call the software-based data centre, that's one big trend. The second is in the storage arena. Our point of view is that flash will be the new disk, disk will be the new tape, and a new form of memory will emerge that is going to be 50 to 1,000 times faster than flash.
Q- It's an area of personal interest for you, right?
A- It's also an area, definitely an area of personal interest. But I think it has tremendous impact in the data centre. I think it will impact the kinds of applications that you can do, much more real time, like I mentioned the real time credit card, fraud detection, when hundreds of millions of users are going after the same thing on the same day or being able to detect when you are digging for oil that there is a potentially hazardous explosion that might happen and being able to get people out of there, a lot of these kind of applications that get driven out of this kind of new thing. So that's the second one. The third area that we talk about is in the area of analytics. And there we continue to move from descriptive analytics to predictive analytics to prescriptive analytics. The other area that we are talking about is in security. There are a number of things that we are seeing in the security area, increasingly the security is going to move again from -- today's security is the way I like - the analogy I like to use is, I wait until my house get burgled and then I change my lock, right.
Q- So it will be essentially built into the apps, the entire ecosystem itself rather than it being plugged in at a later date is what you are saying?
A- And much more forward looking. So I'll be able to, for example, if there is -- let's take an insider threat kind of situation, based on things that employees do, we'll be able to say who are the likely people that might end up being -- doing something bad to the organisation, maybe moving some data outside or something.
We are also looking at privacy, for example, in the context of the Internet of things, which is an important area. But a lot of privacy has to do with regulation too. It's a combination. So, privacy is less purely a technology issue and can be solved purely in a technical way. I think there is technology around privacy, but there is also regulation that has to come.
In the PC business, one of the security projects we have is something we call continuous authentication, as things can get stolen. Today you just type your password once, I could sort of take over your system. Can we change that to continuous authentication based on somebody's writing style, use of keywords etc?
Even in a game, for example, we could measure your physical state, your emotional state and things like that. If I can detect that you are bored while playing a game, can I raise the challenge level of the game so you are not bored as much anymore? In an enterprise setting, it might be that I can detect that you are really seriously working on something today. So if a phone call comes in, I could automatically route it to voice mail and not interrupt you as an example.
Another one could be that, you might be an assembly line worker or a driver of a car or a pilot, if I can detect that you are sort of beginning to lose your concentration level, then... I could get you out of there and save lives or make sure that the machine quality of the product that you are building continues to be at a high level.
Q- Is this a different kind of a Dell in the sense that in the past most of the heavy lifting in terms of R&D used to be outsourced to your partners like Intel and Microsoft?
A- I think there is a couple of things going on. I would say going private is one thing that has changed over the last couple of years. Being in the software business is also a new thing for Dell. That's also happened over the last two years. We've gone from having a zero software business to a $2 billion software business. We're probably among the top 15 software companies in the world. A lot of customers still don't know that. We are on this important journey of going from being more of a hardware-centric company to an end-to-end solutions provider. That is another part of the journey that's been happening over the last couple of years. It doesn't mean that we will not depend on Intel to do the next chips and we'll suddenly get into doing research into the next-generation chip, we are not doing any of that. If you look at my R&D budget, it's got nothing to do with what's the next chip going to look like or even in the memory technology, the way that I'm operating is not trying to build that next generation of memory technology that I talked to you about. I'm scouring a landscape of people and partners and what are they doing and making sure that I'm leveraging the best of what I see out there.
So, it's a little bit different than other companies that might have their own investments in next-generation memory technology, which my experience is that leads to a situation where you are kind of forced to use the technology that you yourself have invested in. If you spent $3 billion dollars generating the next generation of memory, you are not going to look for something else, you don't want to find out there is actually something better than the one that you've invested in. I don't have that limitation. Part of the way we do our research is to scour the landscape, wherever there are good techniques - your earlier question, it doesn't bother me at all that some of our innovation might come from another source. At the end of the day, I'm trying to do the right thing for the customer. I don't care where the innovation comes from.
Q- Are there any moon shots being taken at all?
A- What I have told my team is that I want somewhere between 50 and 70 per cent of the stuff that we work to be successful, 50 per cent is okay. If I'm much less than half, then I think parts of Dell are going to say what are these guy doing, right? And I certainly don't want anything more than 70 per cent, because that means I'm not taking enough risk. So, it's a bit of a fine dance here between how much of what I'm doing is my expectation.
Q- What is India's role in R & D?
A- One of the reasons I'm here is because I want to grow. I'm really looking at trying to grow here in India the next year to have a research kind of outpost in India, leveraging the strength of the R&D organisations that we already have in place in India. My research organisation was started only 18 months ago. And today we are only in the US. We just announced something with IIT Chennai as part of my visit. We will collaborate with IIT Chennai in networking, primarily.
Networking for the reason that we already have a strong networking R&D core in Chennai and there is a strong faculty in that space. It's an important area for us. In India, I'm going to start small. An example would be, let's say, I might have four or five people in Chennai working with a small team at the IIT on a specific project. I want to start to see if this is going to work out. If it works out, I want to do the same thing here, a small team in Bangalore and we haven't gone as far here, I mean, I haven't had the discussion with the Indian Institute of Science or whatever. In India, I would like to work on systems management, and software-defined data centres apart from networking.