 Hi everybody, we're back. This is Dave Vellante with Wikibon.org and this is theCUBE, SiliconANGLE's flagship coverage of the O'Reilly Stratoconference. We're here in Santa Clara. This is our third year doing the Santa Clara Stratoconference. O'Reilly Media does a fantastic job. This is really a show or big emphasis on data science and the actual application of big data. And we're here covering it. This is our third day of coverage. We've been talking a lot this week about security. It's not a topic that when we first started covering big data that a lot of people were paying attention to. They were trying to figure out, what do we do with all this data? How do we monetize it? How do we get Hadoop and MapReduce working and the like? But security has come front and center. We've heard about this all week and we have a segment on security right now. We're going to talk to Eli Khan, who is the Vice President of Business Development and Marketing at Squirrels. Squirrels is a company that we introduced to you last year. And they are focused on security and scalability and performance of NoSQL. Really commercializing the Accumulo project. So we're going to talk about that a little bit and get into it with Eli. Welcome back to theCUBE. Thanks, David. Good to see you again. And we're also joined by David Floyer, who's the Chief Technology Officer of Wikibon. Somebody who's dabbled over the years. David in security, going back to the Rack F days. He's one of the few here who knows what Rack F stands for. So anyway, Eli, let's get into it. We've, as I said, we talked to you last strata. We talked to you at the tug event. You guys seem to be making progress. You're rounding out the team. Give us an update on what's going on with Squirrel and the Accumulo project. Yeah, so it's been a really exciting time for Squirrel over the last few months. Probably the biggest achievement is that we locked down the first version of our product and we've started installing that in a number of client companies. So we're starting getting great feedback. We're installing it both in government clients and also big commercial clients. And our goal is to get a handful of these under our belts over the next several months and then move into production. Yeah, okay, so it's not just the three-letter government agencies that are adopting this. It's more commercial. What kind of industries? I'm getting new information every week about commercial adopters of Accumulo. So as most folks might know, Accumulo was developed originally by the NSA beginning in 2008 and then was open source at the end of 2011. So it was a classified project for the first three years of its life. It's now been an open source project for the last year or so. So not too many people knew about Accumulo outside of the defense and intelligence communities up until about a year ago. I was talking to a defense contractor just this past week and he was telling us that they support 14 different Accumulo deployments within the defense community. He then went on to say that he's supporting twice that many outside the defense community in the commercial sector. So there are now dozens of Accumulo deployments being used across industries, such as healthcare, finance, energy, and life sciences. And the key here is that Accumulo is really the only NoSQL database option where security was built in from the ground up. So if you want to run real time big data analytics on top of mission critical data in production and you're sensitive about security, Accumulo is really one of your only choices. Yeah, so as you say, it's designed in. It's a funny thing about open source projects, right? You don't really know who's using it and you go out there and oftentimes you're surprised by what's going on there. So we heard Intel this week talk about sort of intimate that they might add or try to help add cell level security to H base. Now, David, you've made some comments about that in the past. I wonder if you could sort of weigh in on your thoughts of trying to, what I'll say, bolt on security to an existing database environment. What's your take there? The first thing I would say is that Intel, being serious about security and putting it inside the chip is very good news. That's a good thing, right? We're happy about that. Extremely good news and I hope the chips continue to be fabricated in the US because otherwise if they're fabricated in China we could have some interesting exposures to security. So I think it's a very good thing. I think they've got a long way to go to even put in the first levels of security. Putting in cell level security, that's a great long-term objective but I think it's going to take them quite some time and quite some time to get the performance of it and get really understand what's required at the chip level and how they interface that to the actual application level, the operating systems and the subsystems. There's a long way to go before they get to the stage. As you say, that IBM had in the mainframe is still a long way to go. So Eli, thank you, David. So you're shipping your new product, Squirrel Analytics, but you haven't made a big splash yet so you're taking your time doing the right thing, I think, smart move as a young startup. Talk about, so Squirrel Analytics essentially is designed to simplify the development environment for a chemulo. Talk about what your customers are doing, planning to do with the product. Mm-hmm. So Squirrel Analytics, it's a software platform that enables folks to develop real-time big data applications with much greater ease. So using our platform, no longer do you need to be a PhD level or very sophisticated data scientist to build real-time apps on top of big data. We simplify app development in a few different ways. One, we push down security to the data level. So you don't need to worry about developing complex security requirements or rule sets at the application layer. Security is all taken care of at the data layer. And we do that by tagging individual key value pairs within the database with security labels. We're also, within Squirrel, we're building a number of integrations to enterprise identity and access management systems, such as Active Directory, so that you can connect those security labels to the organizational roles defined in those systems. We're also building a number of APIs that sit on top of Hadoop and Acumulo within our technology stack that expose a wide variety of different analytical capabilities. So we have APIs developed for real-time statistics, for real-time full-text search via Apache Lusine, and real-time graph search. And we've built these APIs on top of Apache Thrift, so that they are polyglot. You can utilize them with any popular web program language. So the demos that we've built for our product, for our Squirrel product, these consist of a couple hundred lines of code in Ruby or Python. And with those simple web programming languages, you can start building real-time apps that sit on top of mission-critical or sensitive security data. So talk about some of the uses that you see your customers applying that they couldn't achieve with, say, some other NoSQL database. Yeah, so some of the use cases involve things like cybersecurity. So the investment bank that we're working with and the telecommunications company that we're working with are really interested in figuring out how to make their security incident event management tools much more scalable. And so these are large multinational organizations that have petabytes of log information. And they want to start combining those log information which are coming from endpoints, could be coming from security network devices, could be coming from web proxy servers. They want to start combining these log file data sets with other sensitive security data sets, such as email information emails that might contain proprietary information or personally identifiable information, might be combining it with things like employee access logs into the building. And by bringing all these data sets together in a single platform and by tagging the sensitive data with security labels, you're able to expose the larger data sets to a much wider population of people. So no longer you have to lock down certain data sets very closely because they contain some sense of information. Instead you can just tag the data that is sensitive and expose the larger data set to all your analysts. So I have a question about the ingest rate. What sort of ingest rates are you able to achieve and how are you achieving that? Is it actually moving towards a near real-time type solutions in this area? Yes, so we can ingest streaming data feeds. So our ingest rates can be in the tens of thousands of writes per second. There's actually a benchmarking study that the federal government is publishing in the next few weeks. It's going through final pre-pub review at the NSA. And in this benchmarking study, they compared a number of the NoSQL databases in terms of their scalability and their ingest rates. And what I've been told by the authors of that report is that accumulo blows all their NoSQL options out of the water, especially when it comes to ingest and scalability. So we're really excited about that. I'm excited to see this report because most benchmarks are benchmarking and so this is an independent source so very interested to see that. We've just got a little bit of time, but I wanted to ask your opinion, your thoughts on this story that SiliconANGLE broke this morning on the Iranian hacker group, the threats they made to the vice president and just in general the threat that is really coming from the world and state terrorism these days. Yeah, so it's a really interesting story. It's something that I looked long and hard at while I was at the White House as a director of cybersecurity there. And while I was at the White House, we were working on legislation that was really going to aim at trying to close up some of these holes in critical infrastructure sectors that these hackers are targeting. So right now it's still pretty much the Wild West when it comes to cybersecurity protection across a number of key critical infrastructure sectors such as dams, nuclear power plants, the energy sector. And what the White House is now trying to do is put out a minimum baseline of cybersecurity standards that can take a lot of the easy attacks off the table. You know, when we talk about cybersecurity, oftentimes it's not the most sophisticated attack that gets people into the system. It is lower level, less sophisticated things that should be easily defensible. So it's scary because there are a lot of holes out there right now and hopefully with some of the White House leadership, we can close a lot of those basic holes. Okay, we're out of time, but thanks for the update, Eli, we'll be watching. I understand you've got some new, you're expanding your team as well, you've got some new management coming on, right? Yeah, we have some great new leaders coming on board, new CEO, new VP of services, a new director of sales. So we have some really great talent coming to lead us to our next stage of the company. Awesome, Cambridge mass based. We love the East Coast angle here in the heart of Silicon Valley. Eli, thanks very much for coming on theCUBE. Congratulations. Thanks for participating. Keep it right there folks, we'll be right back with our next guest.