 Live from San Francisco, it's theCUBE. Covering IBM Think 2019, brought to you by IBM. Welcome back to Moscone Center, everybody. The new improved Moscone Center. We're at Moscone North, stop by and see us. I'm Dave Vellante, he's Stu Miniman. And Lisa Martin is here as well. John Furrier will be up tomorrow. You're watching theCUBE, the leader in live tech coverage. This is day zero, essentially, Stu of IBM Think Day One. The big keynote start tomorrow. Chairman's keynote in the afternoon. Jamie Thomas is here. She's the general manager of IBM's System Strategy and Development at IBM. Great to see you again, Jamie. Thanks for coming on. Great to see you guys, as usual. And thanks for coming back to Think this year. You're very welcome. So I love your new role. You get to put on the binoculars, sometimes the telescope, look at the roadmap. You have your fingers in a lot of different areas and you get some advanced visibility on some of the things that are coming down the road. So we're really excited about that. But give us the update from a year ago. You guys have been busy. We have been busy. And it was a phenomenal year, Dave and Stu. Last year, I guess one of the pinnacles we reached is that we were named with our technology, our technology received the number one and two supercomputer ratings in the world. And this was a significant accomplishment, rolling out the number one supercomputer in Oak Ridge National Laboratory and the number two supercomputer in Lawrence Livermore Laboratory. And Summit, as it's called in Oak Ridge, is really a cool system. Over 9,000 CPUs, about 27,000 GPUs. It does 200 petaflops at peak capacity. It has about 250 petabytes of storage attached to it at scale. And to cool this guy, Summit, I guess it's a guy. I'm not sure if the denomination actually, it takes about 4,000 gallons of water per minute to cool the supercomputer. So we're really pleased with the engineering that we worked on for so many years in achieving these world records, if you will, for both Summit and Sierra. Well, it's not just bragging rights either, right, Jamie? I mean, it underscores the technical competency and the challenge that you guys face. I mean, number one and number two, that's not easy. And not easy to sustain, of course, so you got to do it again. Right, right, it's not easy. But the good thing is the design point of these systems is that we're able to take what we created here from a technology perspective around Power 9 and of course the partnership we did with NVIDIA in this case and the software storage, and we're able to downsize that significantly for commercial clients. So this is the world's largest artificial intelligence supercomputer. And basically we're able to take that technology that we invented in this case because they ended up being one of our first clients, albeit a very large client, and use that across industries to serve the needs of artificial intelligence workloads. So I think that was one of the most significant elements of what we actually did here. And IBM has maintained a lot of, despite you guys selling off your microelectronics division years ago, you've maintained a lot of IP in the core processing and the design. You've also reached out, certainly with open power, for example, to folks you mentioned in NVIDIA, but embracing that alternative processor mode as opposed to trying to jam everything in the dye, different philosophy that IBM's taking. Yeah, we think that workload specific processing is still very much in demand. Workloads are going to have different dimensions and that's what we really have focused on here. I don't think that this has really changed over the last decades of computing. And so we're really focused on specialized computing, purpose built computing, if you will. Obviously using that on premise and also using that in our hybrid cloud strategies for clients that want to do that as well. What are some of the other cool things that you guys are working on that you can talk about? Well, I would say last year was quite an interesting year in that from a mainframe perspective, we delivered our first 19 inch form factor which allows us to fit nicely on a floor tile. Obviously it allows clients to scale more effectively from a data center planning perspective. Allows us to have a cloud footprint, but with all the characteristics of security that you would normally expect in a mainframe system. But really tailored towards new workloads once again. So Linux form factor and going after the new workloads that a lot of these cloud data centers really need. One of our first and foremost focus areas continues to be security around that system. And tomorrow there'll be some announcements that will happen around Z Security. I can't say what they are right now, but you'll see that we're extending security in new ways to support more of these hybrid cloud scenarios. Yeah, it's so funny. We were talking in one of our earlier segments talking about how the path of virtualization and trying to get lots of workloads into something and goes back to the device that could manage all workloads, which was the mainframe. So we've watched for many years system Z, lots of Linux on there. If you want to do some cool container, global Z, that's an option. So it's interesting to watch while the pendulum swings in IT have happened. The Z system has kept up with a lot of these innovations that have been going on in the industry. And you're right. One of our big focuses for the platform, for ZN Power, of course, is a container-based strategy. So we've created, you know, last year we talked about secure container technology and we continue to evolve secure container technology. But the idea is we want to eliminate any kind of friction from a developer perspective. So if you want to design in a container-based environment, then you're more easily able to port that technology or your applications if you will, to a Z mainframe environment if that's really what your target environment is. So that's been a huge focus. The other, of course, major invention that we announced at the Consumer Electronics Show is our Quantum System 1. And this represented an evolution of our quantum system over the last year. When we now have the world's really first self-contained universal quantum computer in a single form factor where we were able to combine the quantum processor, which is living in the dilution refrigerator, you guys remember the beautiful chandelier from last year, I think it's back this year. But this is all self-contained with its electronics in a single form factor. And that really represents the evolution of the electronics in particular over the last year where we were able to miniaturize those electronics and get them into this differentiated form factor. What should people know about quantum? When you see the demos, they explain it's not a binary one or a zero, it can be either a virtually infinite set of possibilities. But what should the lay person know about quantum and try to understand? Well, I think really the fundamental aspect of it is in today's world with traditional computers, they're very powerful, but they cannot solve certain problems. So when you look at areas like material science, areas like chemistry, even some financial trading scenarios, the problems cannot either not be solved at all or they cannot be completed in the right amount of time, particularly in the world of financial services. But in the area of chemistry, for instance, molecular modeling, today we can model simple molecules, but we cannot model something even as complex as caffeine. We simply don't have the traditional compute capacity do that. A quantum computer will allow us, once it comes to maturity, allow us to solve these problems that are not solvable today. And you can think about all the things that we could do if we were able to have more sophisticated molecular modeling, all the kind of problems we could solve, probably in the world of pharmacology, material science, which affects many, many industries, people that are developing automobiles, people that are exploring for oil, all kinds of opportunities here in the space. The technology is a little bit spooky, I guess, that's what Einstein said when he first saw some of this. But it really represents the state of the universe, how the universe behaves today. It really is happening around us, but that's what quantum mechanics helps us capture and when combined with IT technology, the quantum computer can bring this to life over time. So one of the things that people point to is potentially a new security paradigm because quantum can flip the way in which we do security on its head. So you got to be thinking around that as well. I know security is something that's very important to the IBM Systems Division. Right, absolutely. So the first thing that happens when someone hears about quantum computing is they ask about quantum security. And as you can imagine, there's a lot of clients here that are concerned about security. So in IBM research, we're also working on quantum safe encryption. So you got one team working on a quantum computer, you got another team ensuring that the data will be protected from the quantum computer. So we do believe that we can construct quantum safe encryption algorithms based on lattice-based technology that will allow us to encrypt data today and in the future, when the quantum computer does reach that kind of capacity, the data will be protected. So the idea is we would start using these new algorithms far earlier than the computer could actually achieve this result, but it would mean that data created today would be quantum safe in the future. Kind of in your arms race, internally. But it's very important, both aspects are very important. To be able to solve these problems that we can't solve today, which is really amazing, right? And to also be able to protect our data should it be used in inappropriate ways, right? Now, we had Ed Walsh on earlier today. You used to run the storage division. What's going on in that world? I know you've got your hands in that pie as well. What can you tell us about what's going on there? Well, I believe that Ed and the team have done, made some phenomenal innovations in the past year around flash, NVMe technology and fusing that across product line in the state of the art. The other area that I think is particularly interesting, of course, is their data management strategy around things like spectrum discover. So today we all know that many of our clients have just huge amounts of data. I visited a client last year that, interesting enough, had one million tapes. And of course, we sell tapes, so that's a good thing. But then how do you deal and manage all the data that's on one million tapes? So one of the inventions that the team has worked on is metadata tagging capability that they've now shipped in a product called spectrum discover. And that allows a client to have a better way to have a profile of their data, data governance, and understand for different use cases, like data governance or compliance. How do they pull back the right data? And what does this data really mean to them? So have a better lexicon of their data, if you will, than what they can do in today's world. So I think that's very important technology. I would imagine that metadata could sit in flash somewhere and then inform the serial technology to maybe find stuff faster. I mean, everybody thinks tape is slow because it's sequential. But actually, if you do some interesting things with metadata, you could... There's all kinds of things you can do. I mean, it's one thing to have a data ocean, if you will, but then how do you really get value out of that data over a long period of time? And I think we're just the tip of the spear in understanding the use cases that we can use this technology for. Jamie, how does IBM manage that pipeline of innovation? I think we heard very specific examples of how the supercomputers drive HPC architectures, which everybody's going to use for their AI infrastructure. Something like quantum computing is a little bit more out there. So how do you balance kind of the research through the product and what's going to be more useful to users today? Yeah, well, that's an interesting question. So IBM is one of the few organizations in the world, really, that have an applied research organization still. And Dario Gill is here this week. He manages our research organization now under Arvind Krishna. An organization like IBM Systems has a great relationship with research. So research are the folks that had people working on quantum for decades, right? And they're the reason that we are in a position now to be able to apply this in the way that we are. The great news is, along the way, we're always working on a pipeline of this next generation set of technologies and innovations. Some of them succeed and some of them don't. But without doing that, we would not have things like quantum. We would not have advanced encryption capability that we pushed all the way down into our chips. We would not have quantum safe encryption. Things like the metadata tagging that I talked about came out of IBM Research. So it's working with them on problems that we see coming down the pipe, if you will, that will affect our clients. And then working with them to make sure we get those into the product lines at the right amount of time. I would say that quantum is the ultimate partnership between IBM Systems and IBM Research. We have one team in this case that are working jointly on this product, bringing the skills to bear that each of us have in this case, with them having the quantum physics experts and us having the electronics experts, and of course the software stack spanning both organizations. It's really a great partnership. Is there anything you can tell us about what's going on at the Edge? The Edge computing, you hear a lot about that today. IBM's got some activities going on there. You haven't made huge splashes there, but anything going on in research that you can share with us or any directions? I believe the Edge is going to be a practical endeavor for us. And what I mean by that is there's certain use cases that I think we can serve very well. So if we look at the Edge as perhaps a factory environment, we are seeing opportunities for our storage and compute solutions around the data management out in some of these areas. If you look at the self-driving automobile, for instance, just to design something like that can easily take over 100 petabytes of data. So being able to manage the data at the Edge, being able to then to provide insight appropriately using AI technologies is something we think we can do. And we see that. I own factories based on what I do, and I'm starting to use AI technology. I use Power AI technology in my factories for visual inspection. Think about a lot of the challenges around provenance of parts as well as making sure that they're finally put together in the right way. Using these kind of technologies in factories is just really an easy use case that we can see. And so what we anticipate is we will work with the other parts of IBM that are focused on Edge as well and understand which areas we think our technology can best serve. That's interesting. You mentioned visual inspection. That's an analog use case, which now you're transforming into digital. Well, Power AI vision has been very successful in the last year. So we had this Power AI package of open source software that we pulled together, but we made it, we drastically simplified the use of this software, if you will, the ability to use it and deploy it. And we've added vision capabilities to it in last year. And there's many use cases for this vision capability. If you think about even the case where you have a patient that is in an MRI, if you're able to decrease the amount of time you stay in the MRI in some cases by less fidelity of the picture, but then you've got to be able to interpret it. So this kind of AI and then extensions of AI to vision is really important. Another example for Power AI vision is we're actually seeing use cases in advertising. So the use case of maybe you're at a sporting event or even a busy place like this where you're able to use visual inspection techniques to understand the use of certain products. In the case of a sporting event, it's how many times did my logo show up in this sporting event, right? Particularly our favorite one is Formula One, which we usually feature as the Formula One, folks here a little bit at the event. So you can see how that kind of technology can be used to help advertisers understand the benefits in these cases. Got it. Well Jamie, we always love having you on because you have visibility to so many different areas. Really thank you for coming and sharing a little taste of what's to come. Appreciate it. Well thank you, it's always good to see you and I know it'll be a exciting week here. Yeah, we're very excited. Day zero here, day one, and we're kicking off four days of coverage with theCUBE, Jamie Thomas of IBM. I'm Dave Vellante, he's Stu Miniman. We'll be right back right after this short break from IBM Think and Moscone.