 It's theCUBE, here is your host, Jeff Frick. Hi, Jeff Frick here with theCUBE. We are on the ground at the San Jose Convention Center at the Open Power Summit, the first Open Power Summit, which is really an interesting combination of a very old technology and the power microprocessors that IBM's been making for 20 years and now applying open source to that and really bringing a new level of innovation that they didn't have before from an extended and external community. So I'm excited here to be joined by Scott Schultz, Director of High Performance Computing and Technical Computing from Melanox, welcome. Welcome, thank you, thank you. So you're a founding member of this foundation. Why, why is this important? So, you know, it's really important because it really stimulates open innovation, right? Across the industry, experts from all technical disciplines, whether it be GPUs or FPGAs or even companies like mine, we do high-speed interconnect. And we're really the experts in our field. And so it allows to continue the ecosystem and stimulating a more robust, collaborative environment, to continue on to, and what we're primarily interested from Melanox, right, is access scale, right? So instead of rip and replace with new technologies and silicon photonics that you may not need today, because copper is much more inexpensive and much more reliable than silicon photonics has proven to be, right? We believe that there's still a lot of longevity in all of the investments that technology has brought forth up until this point, right? And so as we approach exascale, there are very valid concerns on how we changed software to meet the needs of extreme scale-out computing. But we believe that a lot of the open standards-based technologies today are, you know, with a little enhancement will take us well up into the exascale timeframe before it's completely timed to just rip and replace everything with brand new, you know, proprietary, this proprietary, that nobody likes proprietary. No, right, right. Very much the community is based on open standards, especially in high-performance computing, right? So are you guys involved in other open standards projects? Is this a new thing for you, or is it really just an extension of a direction that you've been going for a while? Right, exactly. This is, it's a fantastic extension of a direction. So Melanox throughout our existence has been very much open based on open standards. You know, the Infiniband Trade Association is an industry standard. It's based on a collaborative environment from a number of technology companies. We're also involved in Open Compute, right? We demonstrated our 100 gigabit a second Open Compute card there as well, and we talked about what's called multi-host capability. And so we're driving new technology, new innovations in these open communities, right, that drive collaboration. And that's really, truly the flagship around what is open power, right? Not being locked into something that is a proprietary type of interconnect, right? Proprietary interconnects over time in the top 500 especially have all fallen off. They've almost all been fallen off. And so open standards and open source and just being open in a collaborative environment is really how I believe that the industry is going to move forward. So talk about the difference between kind of a classic consortium with kind of a defined model versus really what we're seeing more which is open, which is almost more kind of bottoms up. This feels like something in between because you didn't start from zero. We started from an architecture that IBM's had for a long time. But how open is really different as opposed to kind of this, what used to be kind of a top-down standards body that would drive it? That's a good question. So it really is about driving the ISA, right, the ISA behind a power architecture. So what open power is doing though is with enabling the ecosystem around that ISA, right? And being able to provide next generation platform designs, how one might better approach, let's say cancer research or, you know, genomics, anything in the bioinformatics space or oil and gas, right? It really is an industry where everyone's bringing something to the table, okay? It's not all about IBM. It's about being able to innovate around something that is on par with performance and capability and about creating something that is very viable when you talk about cost. You know, if you had $100 billion to go off and build a data center and cost money was no object, you could do it, right? But that's not really how, you know, economics works. So everything has to be price performance competitive, you know, and this is what open power is all about. It's about driving the software ecosystem to enable all of the hardware. And so it's different than a standards body driving a spec just down. This is more collaborative of developing the spec and creating all the software. To support it. For the average Joe sitting on his couch watching this, you want us to know where's the benefit? I mean, what are the main types of applications, the main types of workloads that are really going to benefit from this type of innovation? Well, so that's going to be determined. So I'm not an expert across all technical disciplines. But I do see industries like genomics, okay? Cancer research for pediatric cancer study and this type of thing. Being very practical and very much able to take profit from GPUs as well as FPGAs and all of the acceleration capabilities, right? That's done in kind of this open collaborative environment. And is that because you're delivering performance before at a price that they couldn't afford? Is it because you're delivering performance now at a speed and iteration where they can just move their process faster? I mean, how's kind of the benefit delivered to that person that's working on cancer for children? Right, so it's about price performance, but it's also about time to solution, okay? So we've, unless just take, we'll stay with the pediatrics study. So we've worked with a number of companies in Europe that are doing pediatric studies and by applying RDMA type technology to the process of genomics and cancer research, we've actually been able to reduce the time between, let's say, drawing blood from a cancer patient or someone who needs to have the genomics applied through the whole process flow, okay? What used to take up to three weeks, okay? To be able to develop a treatment, let's say, and to be able to actually apply that treatment, okay, to a patient, we can reduce all the way down into as little as one or two hours, right? So that's the power. From three weeks to one or two hours, right? Absolutely. That's powerful. Right, that's very powerful. And so that's companies like TGen and whatnot are doing research in this area. And so it's not just about, so when you take the industry as a whole and you think about technical computing and Jensen's keynote this morning was robots are taking over the world, it's more about that. When you talk about community involvement and how the open collaboration helps, it's not about doing it one way. It's about doing it the way that's more effective for the solution that you're trying to find, right? So the problems are easy identified, but how you approach the solution could be your way and it could be my way. There's no one right way. Okay. We're both working on the same problem together. And we're trying to be more efficient. Regardless of the walls that we go in every day to work. Absolutely. Well, Scott Schultz, thanks for sharing that story with us. Exciting times. Scott Schultz from Melanix. I'm Jeff Frick. We're at the Open Power Summit, the first one at the San Jose Convention Center. I'm on the ground. You're watching theCUBE. See you next time.