 Welcome to theCUBE. I'm Dave Nicholson, and as part of our continuing series, Does Hardware Matter? We recently joined Andy Brown, Broadcom's senior director of performance and software architecture, for a tour of Broadcom's performance testing labs, where Dell servers with Broadcom components are being put through their paces. Hi everybody, my name is Andy Brown. I'm the senior director of software architecture and performance analysis here at Broadcom and the data center solutions group. And we're gonna take a little tour of our performance lab and I'm gonna tell you and show you where the sausage is made. We're in the performance lab now and I wanna talk to you a little bit about what we've been able to do with these Dell R750 servers. These are Dell's latest generation of Intel Ice Lake big servers. And I just wanna say really how impressed we are with the performance that we're able to get out of these servers and that the Ice Lake processors, the improved memory speeds, the move from PCI Gen 3 to Gen 4 really just opens up the throttle on what you can do with these servers compared to previous Dell generations. And so what you see is we've got the server. It's open right now of course, not typically open to show you what we've got. And one of our Avenger controllers connected into this server and it is running IO connected to this external enclosure over here with 24 drives in it or 24 NVMe SSDs in this enclosure connected to Avenger. And those SSDs are configured as RAID 5 volumes. There are four six drive RAID 5 volumes that Avenger is talking to you right now and communicating with. And with Avenger's architecture and with the performance we're able to get out of Avenger in this configuration, we can see up to 7 million random RAID IOPS. We can see up to 1 million RAID 5 random RAID IOPS which is really unprecedented. And in addition to that, we see up to 27 gigabytes of bandwidth and about 15 gigabytes for the bandwidth in RAID 5 random ranks. So very, very high levels of IO performance and bandwidth. In addition to that, Avenger provides a tremendous ability given its caching architecture to provide very, very low latency operation. If you've heard that SSDs provide low latency which a lot of people have heard that, SSDs provide a lot of latency benefits meaning how quickly it can turn the IO around from the time that you send it to the time that that IO comes back. Avenger is faster than that. Avenger can provide sub-20 microsecond RAID latency given the hardware automation in our architecture and the caching. You know, we talk about cloud. We talk about this server being on the premises in the corporations of the world running their mission critical applications. They don't want to be in the cloud but what is the cloud? Well, to be honest, this also is the cloud. If you go into the aircraft hanger size rooms, there are hundreds and thousands of these and we provide that storage connectivity as well. Our storage is in those servers providing that connectivity that's ultimately presented out to users who need the flexibility and the expandability of cloud-based storage. So not only do we provide the solutions for on-premises, for critical storage applications that need to stay inside the walls of a business but we also provide the storage and Avenger and our controllers provide the storage connectivity needed for today's modern cloud infrastructure. So not only do we provide very, very high level of high-off and bandwidth performance but we provide the latency that really can drive applications, database applications, streaming applications that really crave this type of performance and all doing it in a way that protects your data with RAID 5 or RAID 6 for that matter in a very storage efficient manner in a technology and methodology for protecting your storage and helping around for a very long time that you would trust. So that's what we have right now and that's what we're doing with this Dell R750. We're now on the other side of our performance lab as you can see it's a tad bit noisier over here. This is really where we're gonna take a look at how we test all the performance points that we need to look at with our storage controller and let me tell you there are a lot of those performance points and we need a lot of power and a lot of machines and storage to go do that and we're gonna take a closer look at that here right now. Now you can see we've got racks here. These are full of media. They're full of SSDs, hard drives, different servers and all kinds of different configurations and topologies from eight drives to 16 drives to 24 drives to 120 drives. And we are able to basically measure performance in many, many different configurations. As we know, there are millions of ways you can look at performance and storage even though we can collect thousands and thousands of data points, we're scratching the surface but we have the ability to capture tremendous amount of data to really look at how our controllers perform. And this is just a microcosm of what really a data center would look like. This gets replicated into the large data centers of the world that ultimately allow you to go and access your data through Facebook or Google or wherever you're gonna go get your data. We're here trying to make sure that our controllers and the storage performance is adequate to supply this. This right here represents a small piece of what we can measure and utilize to measure our performance. And ultimately, we've replicated this, as you can see over and over again to be able to in parallel capture data in various configurations, with various operating systems, with various tools and benchmarks to really take a very deep and close look at our controller. And this is what tells us that no matter what you throw at our storage, we can handle it. And we're gonna perform tremendous way to get you that seven million IOPS, to get you that ultra low latency, to get you those ultra high rate five right IOPS. All of this enables us to be able to look at our controller and our solution from thousands of different angles to be able to measure that performance and provide the highest performing server storage solution in the industry. So this is just a small look at what we have and the capabilities we have here at Broadcom to be able to do this. I hope you guys have appreciated this tour and thank you very much. We'll talk to you again soon.