Underneath just about any next generation infrastructure technology there is a database somewhere. But with the rise of 5G, IoT, edge computing, and real time decision making, it’s not the giant centralized databases that are necessarily right for the job. With its integrated data platform, VoltDB has taken a new approach to enabling scalable, low latency, real time decision making for service providers looking to develop services in this new world. With us today to talk about VoltDB’s technology and why 5G could be a huge opportunity is CEO David Flower.
TR: What is your background, and what was your journey to the CEO role at VoltDB?
DF: I am an accountant by profession, but I got into the thworld of technology of software back in the ’80s with the Xerox software division Ventura Desktop Publishing. Then I joined Lotus Development in Cambridge as commercial director for the UK business, back when Lotus, Borland, and Microsoft were the top three software companies in the world. It seems such a long time ago. After that, I started to move into smaller private or recently-gone-public companies including the early GIS company MapInfo for about 11 years before it was sold to Pitney Bowes. Then I got into this world of SaaS with Gomez, which was into website performance analysis and was eventually sold to Compuware. While there, the President and COO of Gomez left Compuware and joined VoltDB, and he asked me if I’d be interested in joining. The company was too young though, at an early stage where money needed to be spent on development and sales rather than senior executives. But we kept in contact while. I rejoined the original CEO and team from Gomez at Everbridge. Then I joined Carbon Black and ran all their international business. That was when I got the call from my previous boss, then the CEO at VoltDB, who said he’d see it out another year but invited me to come over to become COO and then move into the CEO role, and I did. I am not a database expert by any stretch of the imagination, but I’ve always done different things. I’ve always been able to determine good, sensible, logical business models in terms of how to execute the value of the technology.
TR: What are the origins of VoltDB, and what problem was its technology designed to solve?
DF: VoltDB evolved out of the world of academia under the guidance of Professor Michael Stonebraker, a phenomenally well-known individual in the world of database technologies. It was not created to solve a particular business problem, rather it was an academic exercise — kind of crystal ball gazing and way ahead of its time. When the first product came to market in late 2012 or early 2013 memory was still quite expensive. In-memory OLTP databases were all about scalability and distributed architectures. But if you go back to the markets at those times, did they really need that level of scale and low latency? In certain limited areas, yes. But in many ways, it was more a nice to have, rather than an essential requirement. When you’re ahead of your time, that’s what you tend to find, and so VoltDB was trying to find its place in the market. When you are a small early-stage company you talk to anyone and everyone that wants to talk to you. We were found by a few early customers in telco and finance who were looking to improve their BSS application stack. So, VoltDB clearly has its roots as a database, but not as you know it.
TR: How does it differ from our usual concept of a database technology?
DF: When people think databases, they think about long term storage. They tend to manage data as opposed to act on data. By contrast, VoltDB sit at the front end of the data feed. We are not looking to be a long-term storage environment. That is well handled by the back-end systems. We aim to store all that data for a short period of time, but act on that data and undertake a number of complex actions: ingest, store, aggregate, measure, detect, make a decision, and act on that data as it is passing through the application. In the past, organizations would take a number of open source technologies like Apache and combine them to do 60% or 70% of what VoltDB does. And in the days when scale, performance, low latency, were maybe not as much of a priority, they could probably accommodate those needs. But 5G is changing the laws of physics. With 5G, when you’re talking about enormous scale, you’re talking about architecting for smaller footprints, microservices, edge deployment, streaming of data, and phenomenally low latency. And that is exactly why VoltDB was created. We can achieve all of that complex decision making through distributed architecture, at the edge, and we can conform to less than 10 milliseconds to undertake every piece of the action, and with no data loss.
TR: How does VoltDB achieve those low latencies?
DF: When you talk about the world of 5G you are talking about keeping the data close to the processing point and close to the end-user. We are distributed. That is the principle which we operate: no central database or central data warehouse environment. You’ve got to actually be able to act on that data at the edge where it is taking place. You need to keep data and processing connected together and stop pushing it through the network and losing time. But equally, you’ve got to maintain data consistency, which is another unique thing we can do. We just made an announcement in September about our Active(N) solution, which is a multi-active cross data centre replication capability. That enables an edge-data-processing, connected environment without compromising the consistency of the data.
TR: Does VoltDB delve into the actions being taken, or do you just provide a platform for others to supply that piece of the puzzle?
DF: We’re the enabling engine. If you have an ML AI feed you can write stored procedures in there for whatever your application requirement is. Then what VoltDB does is undertake all of that processing and feed that straight back to wherever that action needs to take place.
TR: How much data do you keep out at the edge and how long do you keep it?
DF: We can keep up typically up to about 3-4Tb. The idea is that we keep it for the period of time you need to make that decision or take that action. But you can bring that data back and use it again. It is bi-directional, not a single path.
TR: Other than future 5G plans, what types of applications benefit from VoltDB’s technology?
DF: About 70% of our business is from the world of telco. We are used at about 130-135 mobile operators today, and that’s Tier 1s, 2s, 3s, and 4s. Typically, we are used in their charging and policy systems, but also in areas such as customer value management. We are used in fraud and security, and I think security is a bigger play for us long term. What you’re looking for there is anomaly detection, using AI and machine learning to analyze vast amounts of streaming data and act on it before you’re actually seeing the impact of the offender within your network. And then the other area is IoT, and particularly industrial IoT. We are working with a large Canadian system integrator on a project in the UK for smart meters for electricity to do real-time analysis of energy usage. Similar use cases lie within some of the mobile world in connected car environments. Now, in saying that, VoltDB is not the application, it is an enabling technology. We tend not to deal directly with the mobile operators, we work with partners, companies like Amdocs, HPE, Infosys, Sterlite, and even people like Accenture in the SI space. These are the people that are building the actual platforms or applications.
The majority of the other 30% of our business is in financial services, for example for regulatory performance reporting, which has to be done in real time. We also have some business in gaming and retail. We sit behind some of the world’s largest fantasy leagues. But our focus now is really around 5G. 5G is beyond telco; we’re seeing that starting to blur into other industries such as manufacturing. It’s now spawning brand new applications that could never have been identified in the past, and we see there’s a huge amount of gain for us to win there.
TR: How quickly is the environment around 5G changing in terms of what features and capabilities your customers need from you?
DF: In the old days, telco, 18 months in telco was deemed to be very agile. Now, it’s about days. But the industry still learning about 5G as well as things like Kubernetes and orchestration. We are still really at the tipping point of understanding how people should be actually using all this to full advantage in an operational way. That’s why we work with our partners, such as Amdocs and HPE, who are big drivers in our roadmap. They in turn are driven by their big customers, like AT&T and Vodafone for example. And some of those requests are challenging. But, we are fortunate. We may not have the depth of pockets as an Oracle, but we have an agile methodology. Because we’re focused as well on trying to be the best we can in these specific areas of that technology, we are not distracted by trying to be the best to everyone for everything.
TR: In what ways do you see 5G changing the industry itself?
DF: If you look back at lot of the CSPs operators in the past, they would rarely create anything themselves. They would look at technologies, understand what they need for their business, and then tend to go back through one of the standard partners that exist in the marketplace to do it. With 5G, nobody’s created some of these things yet. They are brand new. So we are seeing some CSPs actually starting to build applications for themselves in these new areas. They’re not going to wait. They’ve got the skills. They’ve got the understanding. And they don’t want to simply be the pipe and give up their networks to the OTT players. They want to take a piece of the pie. Orange, for example, has gone out and created new business units for security and industrial IoT. In these new areas that 5G is opening up there’s no front runner. Everyone’s coming from the same start point.
TR: Thank you for talking with Telecom Ramblings!
If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!Categories: Cloud Computing · Industry Spotlight · Low Latency · Software