 around we've got Paul McKessel here who's the CEO of Clustrix Paul thanks for coming on and Paul's actually one of the co-founders of Icelon systems just got bought out by EMC big data play we heard a lot about that two weeks ago didn't we John yeah I mean you're I'm psyched to have you on because you could talk about a couple different things from many perspectives one you've been an entrepreneur successful Icelon great founding of a huge exit you know billions of dollars old EMC you form Clustrix in 2006 again another startup raise a bunch of money you got clients you guys are very successful great reputation in the industry and you're sitting at the doorstep of this mega trend cloud and growth and applications so you know I think we can talk about a lot of things but you know first thing is as an entrepreneur how do you feel right now in this environment I mean you guys have raised like 30 million dollars you have customers you're producing product what's the state of the market in your mind for being an entrepreneur especially in big data I think there's never been a time better than right now what we have is a situation where folks are trying to build large-scale internet sites that are trying to grow to the size of the entire world for their customer base the challenge about how do you go from systems that were developed to be single server type of solutions to something that can scale across multiple boxes across rows of racks in a data center across multiple data centers how do you do that so it's all about from our perspective clustering scale out instead of having to scale up fault tolerance comes from clustering easy scalability adding nodes into a bio cluster for scalability on the fly or all key concepts things we did at Icelon for unstructured file storage and the things we're developing pioneering and have built at clustering for structured and you built a lot of that at Icelon into the into the software yes I mean that was really kind of the secret sauce and and and so now you're you're a cluster database innovator no so talk about why the world needs another database yeah sure so to date all of the relational database systems that are on the market are single instance kind of things are not truly clustered they're not truly scalable so what happens is your only option for growth is to buy a bigger more expensive piece of hardware get more specialized and get off of the commodity hardware curve and go into something very expensive to remain on these systems the other the alternative to that has been a very intensive development effort alone known largely in the industry of sharding where you essentially break your data apart into individual database systems and you lose all that relational context so everything that the database was doing for you falls apart in this new world so cholesterol for the first time brings the ability to scale out and maintain all of that relational database semantics that has been you know growing with the growth of data over the last 30 years so so tie that to this whole big data theme right because you you think about big data you think about you know shared nothing you think about distributed yeah storage and MapReduce and how do you guys fit and how do you add value yeah that's a great question so related to what MapReduce does for distribution of processing power on large data sets what Clustrix does is we bring the query to the data in a cluster instead of the data to the query so it's very much in that similar theme of parallel massively parallel processing across clustered elements the difference that Clustrix brings is that we are very focused on high performance oil TP so lots of high transaction rates lots and lots of concurrent users in a system that looks exactly like what used to be there but failed to scale so one of the interesting and very compelling parts of our product is that we also emulate and look just like a mySQL server on the wire so there's no need to change your application modify the way that it works do anything special to get scalability and fault tolerance just with Clustrix we slot right in and now you can scale and grow and this thing we were just talking about bringing the query to the data allows you to continually scale out to tens hundreds of nodes in a single cluster with no single point of failure and no asymptote in the scalability and so we have published a couple of different performance results that show linear scalability continuing to add nodes and it just keeps scaling excellent so now so I have to ask you so John was asking you about entrepreneurialism before are you going to hire a CEO or a VP of engineering that's what I want to know oh that's an interesting yeah so my car today CEO VP of engineering we'll see how things program so I'm watching you my kids out there I told you before my kids out there I gotta give a shout out to Roman Roman Volante he's a he's a he's a math whiz he's an engineer I'm like look this is the time to really go hard after the alpha geek stuff right look at this successful entrepreneur it's a math environment yeah math trivets very true and there has been a change I think in recent years you know driven largely by the success of things like my combinator to really push more towards technical based leadership in technical companies and that seems to make a lot of sense and continues to drive innovation and make product fit and company fit align with what people and you're seeing a lot more of that right where you've got technical people now you know coming in and actually running companies you you can't go the other way you can't have a business person go into technology that would be a mess but but totally different disciplines I mean a VP of engineering I mean you focused on products you focus on timelines and prioritization true and and I guess there's some of that obviously in the business side but contrast the skill sets yeah yeah that's a great question I mean I have to give a lot of credit for the navigation of that roadmap right from an engineer engineering background towards you know technology leadership and then company leadership to our our investors or VCs Sequoia Capital US venture partners ATA ventures the folks at Y Combinator people like Don Listwin who's on our boards been very helpful in that and even my co-founder at Iceland Sujal Patel was you know very very helpful and sort of showing me that white combination investor is an investor in cluster access at the beginning or they come in later very beginning yeah okay so you're one of the first classes of the white combination second class yeah Sergei Sergei Sergei sorry off who's my co-founder was at cluster was the one in the second like on later what about the the role of databases now I mean I see with open source I mean anyone can walk into white comedy now they're handing out money for being in class hundred thousand dollars you know so it kind of brings this you know new new collegiate environment to entrepreneurship but you know open source like my sequel and all the you know memcache all these tools they're great I can start stuff I can start a fire and build up with some momentum launching app on iTunes and in launch and have somewhat scale right right you guys address that right so like my challenge is that our blog gets bigger and bigger you know it's all open source it's like shit I gotta rewrite the whole thing I don't want to take it out of production right I don't have fives in databases yeah and exactly that was why we decided to support the the my sequel protocol on the wire so that people can do things like start with a lamp stack right a more traditional open source stack and when you get to the point where scalability becomes an issue clusterics can show up and and slot right in without having to change the application so you can still run essentially a lamp stack but now you get scalability and fault tolerance using clusterics instead of my sequel portion so you guys typical startup you do a lot of good work you'd probably do improve the concepts early on yeah some good tech and good team I'll see biased engineering since you're the CEO sure they got all better bonuses than the other guys but talk about the clients you're announcing you guys just recently announced tuckled here your colleague sure yeah you guys have announced some serious customers yeah that's right now some AOL yeah it's that net box that net photo box and I offer and then we have a number of other ones that that that have become customers since those announcements and we will continue to make and so you get the customers but you had it for a while finally they get the legal signs off on right and you can say things can you talk about them and like what is AOL how are they using because that's not a startup I mean it's still at scale right how are they leveraging clusterics to say a startup that's a girl sure I think I think it's the AOL stuff might be a little sensitive for us to discuss but I can certainly talk about their whole product plans on the web yesterday right sure and and actually you know what I think I would prefer to do is let those guys talk about their use of clusterics on their own blogs because a lot of these folks are starting to they are talking about it they're starting to talk about the not not AOL but the other ones and so we should we should maybe follow up and go back through that so a lot of this is you know technology leadership on their part also right having the foresight to realize where the technology that they had been living with was sort of falling apart not scaling up growing and looking for alternatives and finding clusterics as the alternative that really allowed them to get through their their roadblocks you know it exhibits an incredible amount of foresight and due diligence on their part to even you know find clusterics at the time that they were doing that so let's we were relatively under the radar so without talking specifics in terms of companies can you talk about some use cases and sure help the audience go through a case of their how do you how should they expect to use it what are they going to do to actually bring in and adopt it yeah absolutely so to infrastructure needs to grind typical use case is you've got a application that has some amount of relational data in it so you've got users in your database that are doing things with other users so they might be trading objects or selling things to one another or doing things like having permissions on different objects different videos or photos that they're allowed to see so it's about sharing and information and relationships through the social graph that is a lot of what initially brings folks to to clusterics so what happens is the application is is using the relational database to maintain all of these relational pieces of information and that those queries are usually done through SQL which is one of the interfaces that we support and what happens is over time that database now gets too loaded to be able to be run there's too much pressure being put on the database will be run on a single box anymore and this is why you see folks when they start to grow when their website start to really grow in scale they slow down can't handle the traffic half failures get hung up that kind of thing so clusterics is designed to alleviate all that pressure with really no change at all to the application no change to the way that they work with the systems adding clusterics is as simple as plugging in the boxes turning them on and connecting to it as if it was a database so when you say plug in the boxes you're talking about it comes into the appliance that's correct yes so so talk about why appliance and yeah and take us to a lot of IT guys there's a lot there's real benefits yeah you know there's some concerns talk through some of that dialogue yeah actually Aaron Passey our CTO who by the way was also the chief architect at Icelon has a great post on our blog about why an appliance and I'll just sort of run quickly through that so an appliance allows us to have all the way from the lowest level drivers in the operating system all the way up through the software stack to the very top interface of the database be completely optimized for that piece of hardware completely optimized for performance and even for failure characteristics so what happens when when the storage system deals with a drive failure that kind of thing recovery yeah our system handles all of that and and knows what to expect at different points along the way when disks fail or networks fail that kind of thing so what you get from clusters is a system that understands out of the box its own environment and can react to things that happen in the course of in the course of infrastructure build on the course of life so then additionally there's not any question about you know what kind of hardware are you going to have to requisition and what kind of operating system patches and patches and all that stuff we we take out of all that stuff and just make it very very simple it's designed to be the easiest thing in the world to use install it let it scale and be fault tolerant for you and then you can move on to other things so most of these most of these folks what they really want to do is develop their application add new features for their users and worrying about the database layer is something that should have been solved and hadn't been and so we really attack this problem based on the failure that we saw in the industry to address so it's another point of management but it's just that it's one point of management as opposed to all these bespoke that's exactly right great so just on cluster x where do you see the cluster x in the next five years I mean you see it what's your vision I mean you have see started it because you had a passion for this whole database and growth side of its scale what do you see in five years so vision yeah I mean so I so long as a company that we started for unstructured data for file storage you know we went public after five or five and a half years and then emcee made a very great you know intelligent purchase on their part in December bought Iceland for just over two billion dollars I know you're biased but I agree by the way I actually think it was a great purchase yeah so I'm really really smart and in the unstructured data space that made a lot of sense so in the structured relational data space we see cluster x doing the exact same thing we have by many analysts opinion a much larger market in the relational data space than in the file space more intelligence that needs to go into the query planning of relational database queries as opposed to simpler file queries and so so there is I mean people are mashing up the database environment who go like say Facebook for example they do a lot of things with the photos of the haystack projects from well-known about so the system software is a hack in a way but now it can't be a hack you need to have the scale out mentality around the structured that's right not just the unstructured but like that's right the real production that's right and and you know we see ourselves as preventing people from having to go through the kinds of growing pains some of the earlier pioneers had to go through some of the folks that they're right there on some of the folks at Facebook they had to do incredible you know feats of engineering to be able to get through that stuff through those growth growth phases and we see ourselves as enabling other folks to be able to get through that without having can they still code can they still code so you save so you basically save a lot of time benefit to the your client and risk and and diversion and all kinds of things we say them and so can they still code with your system so if I want to do a little bit of Facebook hacking like environment like for my environment we still do that sure absolutely yeah absolutely the the just because the database is now scalable doesn't mean that there aren't more interesting things you can continue to do on top in the in the application stack what's your assessment of big data I mean we ask all the CEOs that come in here what's big data I mean everyone's it's like web 2.0 yeah everyone's got a different definition what's your your take or someone sees you the cocktail party yeah you know tell me about this big data thing what's going on with right when we've talked a lot today about the internet sites because that's our initial focus but beyond that even in a lot of the more traditional enterprises and and things like medical research genomics in a lot of the places financial markets things like high-frequency trading right you're talking about being able to capture and manage as much data as you can because it's not just the individual pieces that have value it's the it's the aggregate in total picture of what is happening in whatever segment you're tracking that adds the real value right so that is why and we're reaching the point now with the size and speed of these computer technologies that we can actually hold that in our hands and manage it and control it right so that is where big data is coming from that's where it's continuing to go internet is just one space out of the entire swath is really being enabled by where things are at so what's going to happen in your mind final question Mike at the top of the stack is I mean ultimately virtualization all these kind of plumbing activities enable the next layer up database middleware sure we're simplifying it but but that's what you're providing you're enabling up scale yeah so and data is a big part of that and center yeah we're trying to let people get out of the business of worrying about their own what's going to happen update the stack applications I see iTunes is a great you know app stores a poster child for growth angry birds where people using it business is gonna do that what's gonna what's your vision for the next you know five years or so in the top of stack virtualization at the desktop and application integration what's your angle on that yeah sure I mean we see lots of different things with so for example dealing with file storage in an inter-corporation how do you pass files back and forth between your different offices make sure they're tracked secure all that kind of stuff was traditionally an in-house IT managers thing to deal with right there's in some cases the real bane of their job right but now with folks like a box net providing secure file storage between folks and focusing on enterprise as a large part of their target we see we see people being able to sort of out this is IT as a service yeah exactly IT is exactly right IT is a that's a big trend you see that done deal absolutely you see that happening absolutely yep cool okay we are