 from the MGM Grand Hotel in Las Vegas, extracting the signal from the noise. It's theCUBE, covering splunk.com, 2015. Brought to you by Splunk. Now, here are your hosts, John Furrier and George Gilman. Okay, welcome back everyone. We are here live in Las Vegas for splunks.com for a special presentation for theCUBE. This is SiliconANGLES flagship program. We go out to the events and extract the signal from the noise. I'm John Furrier, the founder of SiliconANGLES. I'm John Michael, George Gilbert, our Wikibon research big data analyst covering the space and our next guest is Raghu Nambar, who's a distinguished engineer, chief architect of the Big Data and Analysis Analytics, covering Cisco. I mean, I'll see you there in Big Data. Welcome back to theCUBE. Good to see you again. Yeah, good to see you. I think this is the third time I'm in theCUBE. Really excited to be here at Splunk in Las Vegas. We'd love having you on one because you get to share some really good data with us on what's going on. But also on the engineering side, there's a lot of folks out there right now re-architecting their networks. They're re-architecting this full stack of what DevOps is, Big Data, Analytics, the whole dream of the modern shift to this new modern era. So, you know, it's exciting, right? Yeah, absolutely. If you look at the UCS portfolio, right? I mean, we talked about this one in the past interviews. We created the portfolio for a new generation of applications. So right now, you know, we have a UCS Blade Systems, which is the most popular for enterprise applications. Then last year, as part of our UCS 2.0 launch, we have announced two products. One is a UCS M-Series for high compute density applications and also for test and development applications. UCS M-Series, in two RU space, we can pack 16 servers. In a two RU? Yeah, 16 servers. And so we have a unique technology called the single link in this architecture. So all the servers share the same set of components, the storage and the network. When you say just to be techie for a minute, when you say 16 servers, I assume you mean 16 sockets. No, I'm talking about the 16 independent servers. More like 16. Some of which are dual sockets. So we have the option of single socket as well as dual socket. So it could be... This server means a server has its own operating system instance. So it could be 32 sockets if you chose. No, actually it is 16 one socket or four two socket. Okay. And it is the M-Series. And another platform I want to highlight is the C3000 series. So this is a dense storage platform. Okay, in four RU space, you have the option of one compute blade or two compute blades. Both are two CPU and up to 60 large form factor describes. Okay, when you talk about 60 large form factor describes with eight terabyte storage capacity, you're talking about 480 terabyte of storage capacity in four RU. Put 10 of them in on a 40 to your rack. And now you're talking about close to five petabyte of storage capacity. So on a platform perspective, we're ready for two days big data and tomorrow's big data. So I want to get into the big data. Certainly we'll see you guys next week in big data NYC event that's happening. And then of course strata had due for all going on as well. But I want to talk about this event here for the moment for a second. You guys had a benchmark up on stage so I met Kartik and the team. We had a customer on earlier here on theCUBE that was just like blown away by the performance. And they were like, speed always wins. Customers love speed. Can't go wrong with big things. Go faster. Talk about that benchmark. What does it mean? Physisco customers. Yeah, absolutely. We have been working with Splunk for two years. This is very strategy partnership. And we have announced a set of reference architectures and performance benchmarks last year. Okay, now I'm sure like you have heard of the keynote, the major announcement here is Splunk 6.3, okay? When you do a performance comparison, 6.2, the previous version versus 6.3 on a like a generic platform, I believe they have seen like two X performance improvement. What on Cisco UCS? We have seen four times performance improvement for indexing and about six times performance improvement for surfing. Significantly better than previous generation of, you know, software. We have seen this kind of performance improvement with the hardware innovations but from a software innovation, it is pretty exciting. So now, you know, you'll be able to provide four times more processing power in a box, right? I mean, typically what we have done to get more performances like scale out. We definitely continue to support scale out. It's going to be a big part of the portfolio but here the ability to pack more processing on a server basis. So I got to ask you about the Cisco value problem. As you move into Splunk for two years, great endorsement for Splunk to have a partnership with Cisco and vice versa. Splunk is no longer a startup anymore. They're a public company. They're one of the big guys at the table within IT. Cisco install base is huge with Enterprise. You guys are the gold standard for routing and networking gears now with UCS and the data center is really unifying everything. Yeah, absolutely. So what does that mean from a Splunk standpoint? What is Splunk doing today to give your customers new insights and value and what new value is being created on top of that? Sure, I mean, you know, as you said, Splunk has 10,000 new customers. If you look at UCS, so we have about 50,000 new customers you know, since we introduced a product in five years. Okay, what Splunk customers get from Cisco? I mean, the main differentiation that they get from Cisco is the UCS differentiation, the ability to provision servers, okay, manage and monitor servers and of course, high performance that we have demonstrated through these benchmarks. And in addition to that, we have created Cisco validated design. We call it CBD in short form. This is basically an end-to-end deployment guide for our customers to deploy Splunk on UCS. You know, I mean, if the guidelines and the CBD is followed, we guarantee out-of-the-box performance. And also like, you know, it's very prescriptive 300, 400 page document. Yes, all the information that you need, including how to scale from a small configuration to very large configuration. And in addition to that, we have highlighted one more important thing for data lifecycle management. Okay, you have your hot and warm data on one tier and most of the processing is going to happen over there. But you have to, you know, move the data to a less expensive tier after data get older, right? So we have this C2000 based platform, ideally built for that. So here it's performance. Here it is capacity or dollar per terabyte. This aspect is also articulated in our Cisco validated design. So I'm going to ask you from a big data standpoint, we've always, this is our fourth year covering Splunk.conference. And it's been a progression every year, but it's always been, at the heart of it, a big data opportunity. But yet it's driven out of the geeks and IT, which you guys live in that world as well, one part of your business and growing now to go to the edge of the network with Cisco. As you guys look at UCS unified as a word, that Pat Gelsinger at VMware used in his keynote, he's calling it unified hyper cloud, which is interesting because Jim McHugh was on our crowd chat saying, hey, we love this word unified. So what does that mean to you? What does unified mean to you from a customer standpoint? As you look at the picture, in context to the big data opportunity, because you've got internet of things, internet of everything from Cisco standpoint, just basic IT walking and tackling. There's now the confidence of those two things coming together. So unified, I think there are two unifications here. The U in UCS, Unified Computing System, that U basically unifies compute, storage, compute connectivity and the storage access through your unified fabric technology and the ability to manage entire application cluster from a compute, storage and connectivity perspective. IoT is the next big thing. We have been in this space for a long time, even though it is called IoT, we just started calling it. Well, Cisco had that human network advertising, you guys know the edge of the network, it's been there forever. The whole idea is, you know. Connected device is a device, IP address. I mean, you know, if you look at today, most of the analytics are happening at the core. I would say like 80% of the analytics are happening at the core. But that's got to change for IoT. Absolutely, I mean, if you look at five years from now, not even five years, three years from now, I would say like 50% of the analytics are going to happen at edge, because the applications are demanding SLAs and response time, you know, right when the data is produced. I mean, send the data on the core and running analytics, there is a latency. And the compute is increasing at the edge too. From whether you look at your iPhone or device, you can put compute or move compute there, right? So we have a set of products that is already there and we'll have more innovations, like intelligent routers and devices specifically built for edge analytics. What is the number one thing? I want to get your take on this. As an engineer, why is UCS been so successful? I mean, I remember when you guys launched UCS, a lot of the naysayers, again, we do the QBIT, IBM, HP, Dell, you all the events we do, you guys had a lot of people coming out of the loopers, oh, UCS is never going to work. IDC had market share numbers that was slicing and dicing. Everyone's jockeying for we're running in service. What has been the reason for the success of UCS? There are two things. One is the innovations in UCS, okay? Unification of fabric, that brings simplicity and significantly lower cost. Second thing is the manageability and monitoring. Again, UCS manager is bringing on the table, significantly reducing the effects, okay? I mean, the second thing is the performance. So we have consistently demonstrated with industry standard benchmark, Cisco is number one in every category that you can think of. This is like beating our, I have great respect for our competition, but the competition that has been in the business for forever, right? I mean, these are the two main things, you know, that is, but it has helped UCS to grow into a 3.5 billion dollar business in five years. But there's, I mean, there's got to be something, some special sauce that was unusual because like Oracle, I'm closer to their engineered appliances and, you know, they use Infiniband, it's not too proprietary. I mean, they baked a sand and a database cluster together, but you guys put this, you know, sort of special fabric that UCS, you know, Cisco's specific and Oracle's really struggled with Exadata and their whole engineered appliance business. They've struggled? Yeah, they have. You wouldn't know what's coming up. I don't know. I've seen the numbers, I've seen the numbers. I challenge that. I think, okay, go ahead, continue, but. So, if you look at the fundamental networking in Oracle Exadata or like any traditional platform, you have at least three types of connectivities. You need typically one gig connectivity for management, managing all your servers, you know. Second is, you need like a 10 gig or 40 gig connectivity for your cluster to connect, no to talk to each other. Third one is storage access. If you want to bring in data from external storage, we are talking about three different types of cards, three different types of controllers and three different types of switches. In Cisco, we unified everything into one unified fabric, one card, one set of wires, okay? And unified fabrics. Right, give us an insight into what the architecture is for the customers out there with Splunk and Cisco because obviously you can help Splunk with your 50,000 customers, they have 10,000, but also they help you because they offer a great value proposition for your customers. But what does that architecture look like? What is the, in the mind of the IT architect and or potentially software or DevOps environment, which is more IT off from this standpoint, from this vector, what does that architecture look like? So what we have done is we worked closely with the Splunk engineering team to create a set of reference architectures. The reference architectures are based on our experience and experience from our Splunk engineers and also like feedback that we got from our partners and customers. So we created two set of architectures. One is for performance optimized with our C220 servers. So basically that platform can support 250 gigabytes of indexing capacity per day with a one month retention. That is performance. And we have also created a capacity optimized SKU for where data retention is critical. It's based on our C240 server, 80 gigabytes per second indexing capacity per day with a one year retention. And these two configurations can be deployed as a single server or a cluster of servers as the workload demands. We really appreciate you coming on theCUBE, sharing your insight. Now we're going to probably talk to you guys next week at Big Data NYC. But I also got to say that Cisco is not only innovating, bringing the content here on theCUBE and in our kind of open source framework that we call theCUBE, but they bring in now goodies. So we have all kinds of good gifts here. So the new standard is come on theCUBE and get the Cisco hat, get the giveaway here. I got the Cisco pen, the Cisco Chachki here from my iPhone, but also the special SanDisk wireless thanks to Renee Yau for doing the gifts, connecting us all together from smart cities to UCS and the data center. You guys are redefining the edge of the network. So give you the final word, share the folks out there. What is the new mojo for Cisco? What is the new Cisco all about now in this new modern and infrastructure? So I mean, Cisco is all about the connecting people. Okay, and we want to power the next generation of enterprise applications, cloud and internet of things. Awesome, thank you so much for sharing your insight. It's theCUBE day one coverage live here at Splunk.conference. I'm John Furrier, George Gilbert. We're here with Cisco in theCUBE, breaking it down from the distinguished engineer himself, Raghu. We'll be back with more live coverage after this short break.