 Patel welcome to the Cube. Good to see you again. Good to see you. Remember we did a podcast with Chris and down the downstairs conference room. Absolutely. Congratulations. Welcome to the Cube. This is our flagship telecast where we go out to the places where there's action happening. Also here we're exclusive here with HP's moon shot announcement. Chandra Khan you are the guru. Now your official title is Senior Fellow Director Sustainable Ecosystems Research Group. But we know we've talked before. You're a big part of it. This is my co-host Dave Vellante. So you're pretty excited. I mean this announcement is right in your wheelhouse. Absolutely. So just tell us how do you feel about all this? I think this is a key stepping stone towards what we are trying to achieve in terms of using the IT ecosystem made up of billions of handheld devices, thousands of data centers and digital print factories to deliver a net positive impact while making money for the company. So having this kind of infrastructure element enables us to use IT for resource management at large scales of the scales of cities, which is a big opportunity and it's also a big opportunity from a sustainability perspective. So when Prith Banerjee took over HP Labs I think two years ago maybe two and a half years ago sometimes he got eight initiatives. I think it was about eight initiatives I think. He wanted to solve big problems. We talked two years ago around data center, energy, follow the sun, all the stuff that we talked about was really great conversations actually happening now. One of the things HP has been criticized for as a company is oh the labs is amazing, they have great people in the labs, they just can't get products out the door. This is actually not the case. This product announcement today was really an example of a very fast HP. Absolutely. You have HP Labs, DNA, you have a division shipping a product, ecosystem, whole new generation. So talk about this shift. One, HP internally, how you feel about that and describe that and then to talk about the second generation technology. Yes, so first HP internal as you said. If you see the way we have done our work we've established thought leadership, advance the state of the art. So when Prith took over he emphasized that we need to be externally very visible to show thought leadership and establish and show how we have advanced the state of the art but equally important or even more important was to deliver business impact. So the way we have gone about doing it is we have presented our thesis shows how the thesis advanced the state of the art, got collaboration from our peers externally and then we have gone on to build demonstrators. So as an example the area where I lead we have building a demonstrator that shows the data center in which these systems will live and the data center runs on its own supply side microgrid and we are driving towards a net zero data center. What we consider the blueprint for the next generation data center described net zero data. So the way we describe it is as follows. We use available energy or exergy from second law of thermodynamics as our driving principle. So an available energy is energy that is useful. So if you take diesel as an example 35 mega joules per kilogram once you burn that diesel that availability is gone and it's gone forever. So we take a supply side perspective given a pool of available resources diesel, sun, wind. How would you locate your data center particularly in geographies where grids are not available? How would you build your own power plant consisting of sun, wind, you know maybe non-renewable and in some cases other baseloads like manure, biogas, use that microgrid, use your data center with systems like the one we are announcing and then service SLAs that are emanating from all over. There are people who want that would have service level agreements that require have an interactive service level agreement that requires immediate attention. Then there are people like our colleagues who do animation who want batch loads to be executed. They want it to be done in a given time and then there are billions who have not boarded the IT bus who want IT services may not even have electricity at home. You know they want IT services to better improve their life like a vegetable vendor who can schedule you know get orders pick the right vegetables he doesn't have fridge at home he wants IT services useful IT services. Man has no electricity at home. Do you think he can wait for the sun to come out? So if you build the data center with the microgrid with multiple sources of power you take all these service level agreements with variety of workloads then why can't you dynamically allocate them so that you shape the demand such that you are below the supply side curve. So if your supply side made up of sun, wind and other sources gives you a given supply of energy. If you take all the demand coming in and you shape it and you are below the supply you are net zero. That was a long explanation. Local farming concepts brought to the data center. Absolutely. Local farming concepts. So what is your blueprint then for the sustainable data center? Is it that? Is it more comprehensive? It's quite comprehensive. It's a holistic perspective. So the blueprint like we said is built on a technical framework on available energy. First principle exploit local sources of available energy in the line of being more comprehensive. Second, examine waste, available energy in waste, waste stream. So when I'm driving by a sugarcane plant in my village near India, I see flue gases coming out at 500 degrees Celsius. One joule has 0.5 joules that's available. You can harvest that available energy. That's the second principle from a supply side. Yet another principle is take the life cycle energy used. Available energy used across the life cycle to dematerialize our data center. Minimize the amount of material used in the data center. So energy used in extraction, manufacturing, operation, and reclamation. So we build tools for that. So on the supply side the principle is use local sources of available energy, use availability in waste streams, do life cycle energy management. On the demand side, give you what you want when you want it no more no less. So given this pool of available energy, how do we broker the service level agreements? How do we allocate? In that context we have developed control systems that dynamically schedule cooling, power, and IT. So use virtualization to move workloads around, adjust the air conditioners based on temperature sense in the data center, and adjust the power loads. So this comprehensive view of what we call end-to-end design and end-to-end management is the blueprint of the future data center. So the devices they announced today, obviously the products that fit into that, but there's other elements like software, probes, and things like that. That all fits under the blueprint. Absolutely. Under end-to-end management blueprint. So if you think of having these devices, so the devices we announce have flexibility in them that we can exploit with a layer that is a knowledge discovery or analytics layer that aggregates all the data to schedule things internally. Then there is a control loop. All these are from a management perspective, the software that you guys are at HP Lab, you're always inventing the future at HP, and the hot thing in the market today is big data analytics, real-time analytics. This is not new to you. So how do you see all this playing out with big data and analytics? Fundamentally, instrumental to the design of the data center because you need to have those analytics. What's your view of big data and the possibilities? So you will hear a lot of, from my view, is very resource management centric. So you will hear a lot of viewpoints on using big data to look at consumer patterns and consumer buying patterns. Those are all well and good. My view is an error that we have just started working on, built on this IT ecosystem and the data centers. There's a huge opportunity to mine structured and unstructured data from physical systems. So under the manhole colors, in the pumps, in the compressors, in the highways. So what we see a big opportunity with big data happens to be management of resources, physical resources, power, water, waste at the scales of cities, at the kilometer scale. So imagine San Francisco of tomorrow, where we are getting data from that big inner pipe that is traversing a pipe to see if there are cracks in the pipe. Imagine the data that's created with that. And imagine the usefulness you deliver with that. So clearly we see a huge opportunity. So you're actually instrumenting those industries? So what we have started is in, as the need arises, for example in this building we have instrumented the building. So we are gathering the data to do anomaly detections to see when things are working inefficiently, how to make them run efficiently. Those kinds of fillers have started and cities are starting to look at using these kinds of approaches using IT. Yeah, so we've had Dave Thomas Alley on before he's talked about sea of sensors. Correct. Right, so this is what we're talking about. The sea of sensors that Dave was talking about is the sensors we use in the systems, in the data center to dynamically manage resources in the data center. Now take those sensors and scale them up kilometer scale. Or for matter, not only sensors, already if you look at a city and somebody doing maintenance in a city, they're creating all kind of unstructured text in notes they take. There is all these devices like compressors and pumps that are already, shall we say, tweeting their state. So if what I'm saying is let's harness the tweets of the physical systems. Let's use IT to manage and say, Mr. Pump, how are you doing today? We don't have any more money to buy a new more pump. We don't have money to build infrastructure. By golly, we're going to make the infrastructure last longer, given the economic crunch we have. And then there are other parts of the world that are saying, we are building new cities. Help us build it. So, and you have alluding us, this is fantastic conversation. Love this and talk for hour. But, you know, we now have, this is how the Internet of Things, we have mobility. We have wireless. We have different types of wireless networks. What areas do you see that are fascinating right now to you around expanding this thesis to being, not just demonstrations, but actual deployments? Is it wireless? Is it mobility? Because now you have mobility. You can actually have someone capturing an iPad and not even have to go to the pump. Yeah, yep, yep. And then only going when it's efficient. Right, right. So is it the wireless? Is it the mobile? Is it the cloud? All the above? Is it a confluence of everything? So you hit the nail right on the head. Internet of Things, cyber physical infrastructure is, you know, increasingly what we have today. And it's all of the above. I can't believe. And by the way, it won't be an iPad. It will be an HP laptop with the ability to do computation. So, you know, I feel that in this whole world where we have gone to all these devices, we have become very much centric on consumer type apps. I am from the generation that does finite element analysis, CAD design. My laptop is a workstation. So an HP 1000 used to manage industry, industrial processes, right? I think that we are going back to an era where laptops will go back to what they used to do best. They managed buildings and resources. Because a lot of things need to be done in real time. Yeah, it's a mini computer mainframe in your hand. Exactly. And so my feeling is, if you're managing a factory like a pharmaceutical, do you have time to take the data, send it someplace, analyze it? You need to do real time close the loop. And you infrequently dump to the cloud to do analytics of historical data. So what I feel will happen is real time control will happen with a laptop workstation or a handheld device. And historical analysis will occur via cloud. And the handheld device will also be a sensor. What are the implications of that for software quality? I mean, we've got dozens or sometimes hundreds of microprocessors in our vehicle, buildings, and if I have to reboot the system every other time. The complexity issue, really. The complexity is amazing. I think that's a very good question. And I suggest we go back to old school engineering. So when I worked on this drives, we used to be very cognizant of the kind of stuff we put out, right? It was not like, okay, we'll pay to test it on a bunch of people. When you are just doing some things that are not going to shut down your city, electricity, then it was, it is fine. So the Google car that's in beta is not a good car? Well, it's a great start. That's starting in 101. No, it's a great start. I mean, that's a great start. But you're saying, good enough is not good enough. Well, don't test on our people. So what you would do is you start with that, and that's a great step forward, right, in showing an example. Then you go ahead and test it out like we always did when we did physical industries of the type we did and where HP1000 was used to manage our plants, right? We are going to have to do that. So that, you know, what recently that incident in Yuma caused entire San Diego to go blackout, right? So when IT will manage cyber physical systems, IT tools and software need to be very robust. And we need to go back to old school design and engineering. So let me ask you just kind of more of a kind of come out of the HP labs, put your, you know, HP employee industry person hat on this announcement today. Talk about it from your standpoint. What's the most important thing for folks to take away from this announcement about the server technology that they're using with the ARM technology and the power for the data center? Yeah, sorry, I get excited. So you're saying, what you're asking is the way I see it is you must take this announcement in the holistic framework that is HP, right? We are not announcing when, when, so you are talking to me yet the people who have done all the work are some other people who will be talking to Partha recently. But when you look at baby, that's his baby. So, but when you look at, so Partha will be talking after this, but if you look at it, we work with, you know, we all work together. In not only designing an efficient server, but how does the server serve as a flexible building blocks? So it goes into the data center and that it can be managed in the data center efficiently, dynamically, so that we build the net zero data center. And how do we use the net zero data center to manage cities? So the key takeaway for people is that HP is rich with mechanical engineers, electrical engineers, computer scientists, economics, who work together in a multidisciplinary fashion to create a building block that becomes a solution of tomorrow. So it enables a solution of tomorrow. And without the building block, we don't get the differentiation. So you see this as a differentiation for building unique cloud services that we are used to and new cloud services of the type I described that will manage resources at the scale of cities. Okay, Chandra Khan. Thanks so much for coming in the Cube. Great to see you. Chandra Khan Patel. Always great. Google here at HP Labs runs the research group here about sustainability. And thanks for coming on the Cube. Appreciate it. Great to see you again. Thank you.