Sramana Mitra: You’re telling me that your go-to-market strategy is OEM?
Don DeLoach: I would say it’s more and more OEM for sure. It had been a combination of direct sales mostly into the ad tech space and OEM sales into the people servicing mobile network operators. Without doubt, that’s over half of our business right now. It is asymptotically approaching 100%. The interesting thing there is that the mobile network operators are really at the forefront of driving the Internet of Things. When you step back and read the various industry analysis or some of the reports coming out of Cisco, they project 50 billion connected devices out there by 2020.
The fact of the matter is these are devices that are producing tremendous amount of machine-generated data. If your specialty happens to be analyzing and storing machine data, then this trend looks pretty good. We are focusing a lot of our attention on that space both in terms of the telco space as it exists today and the challenges they’re confronting with the increased volume that the providers have to deal with. By extension, we are focusing where we’re taking this technology in line with where we think the Internet of Things will drive the market.
Sramana Mitra: I’d like to take the two big use cases – ad tech and the telco use cases. Take us through specifically what’s happened in the ad tech and the telco world.
Don DeLoach: In the ad tech world, we are used for a variety of use cases across the overall ecosystem of the ad tech space. That would be everything from supporting parts of the real time bidding process to configuring campaigns or tracking productivity of the ad campaign. A lot of these firms are 60 or 70 people and they are growing at magnificent rates and they’re struggling to keep up with both the amount of data and the burgeoning business that they have. One thing that is typically in short supply are technology resources for maintaining things like databases and the IT infrastructure because they run very lean. One thing that I haven’t spoken about which I probably should mention is that the unique capabilities we have are extremely important in environments like these because of the human capital.
Let me back up for one second and just highlight what some of the unique capabilities are and then I’ll apply those to these use cases in ways that it probably will help make sense. Infobright fundamentally allows you to ingest very high volume of data at very fast load speeds. We have mobile network operators that are ingesting two terabytes an hour and more. What we can do is accommodate that in a way that establishes this metadata layer. This is where the underpinnings of rough set mathematics come in. When we ingest the data, we do it very fast and we build a metadata layer and the focus of that is to capture the maximum amount of information about the overall dataset and store it in the minimal amount of space. Then, we further compress raw data at about a 20:1 ratio depending upon the data set. Sometimes, it’s actually significantly higher. On average, it’s a little over 20:1.