Main content

DataStax CEO - ‘We know our tech solves the use case, but there is a skills problem’

Derek du Preez Profile picture for user ddpreez April 19, 2016
Billy Bosworth explains why DataStax should be the solution of choice to power enterprise cloud apps, but admits maturity brings its own problems.

Billy bosworth datastax
Billy Bosworth, DataStax CEO

The NoSQL market place is becoming increasingly competitive and is rapidly evolving. It’s a complex landscape to get a firm grip on. For instance, NoSQL doesn’t cover just one technology type, it’s a diverse mesh of open source tools that a number of players are quickly trying to pull together, make manageable and introduce enterprise capabilities.

One of those leading players is DataStax, which has its beginnings in the distributed Apache Cassandra architecture, which operates as its core engine. However, since then it has begun building out its architecture by integrating with Spark, Kafka and by acquiring Aurelius, the leader in TitanDB graph technology.

During DataStax’s European Summit keynote this morning, CEO Billy Bosworth explained how DataStax’s core engine is coming together (see picture below) and is fulfilling use cases for companies that want to take advantage of more sophisticated internet/cloud based architectures. Largely, he sees the use cases in personalisation, getting a 360 view of the customer and Internet-of-Things.

However, at a top level Bosworth wants DataStax to be known as the company that can solve the database problem for enterprises that are trying to knit together these open source technologies together themselves when building cloud applications.


I got the chance to sit down with Bosworth at the event, where he explained:

We really are focusing on this notion of helping enterprises build and run their cloud applications. And the great part about that is, if you ask the customers, they will tell you that their biggest skills gap problem is on these cloud apps and these distributed apps. That is where they are struggling.

The product evolution might look like that it’s getting more complicated, but really it’s a simplification of what they’re trying to do on their own, inside their own companies. When you look at it, it’s really getting crazy the tapestry of these open source technologies.

Our whole mission is to bring you a new data layer that can handle all of the demands of these cloud applications in a single offering, so that your ops team can get their hands around it. Your security officer can get her hands around it. Your developers have one model, one access model they can get to and then they can model their model data different and handle all the things in one system.

For us, it’s about the simplification of the back-end for these cloud applications. We have to make that easier.

Bosworth explained that what he sees happening in the market is a growing disconnect between lines of business inside enterprises and their IT team. This isn’t unusual and is something that has become increasingly common with the onset of cloud.

He said that line of business developers are being told to go and figure out how to build new innovative cloud applications, underpinned by these open source technologies, where they end up trying to stitch them all together themselves. This model is then replicated across a number of lines of business within the same company, without IT ever being involved.

Bosworth said:

Right now we are at a stage where companies have been solving it one line of business at a time. And one application at a time. Now organisations are realising that they can’t scale that, they can’t handle it from a security standpoint, they can’t handle it from an apps stabilisation standpoint, they can’t handle it from a developer training standpoint and they can’t handle it from any ROI perspective.

We get it, you totally need the power of these open source projects. You don’t need ten open source projects, you need a core back-end system for your cloud apps. That’s what we can do for your database.


The big news for DataStax at the moment is its tighter integration with the graph technology since the acquisition of Aurelius. Unlike Cassandra, which is useful for streaming data in real time, and unlike Spark, which is useful for analysing repeatable queries in real-time, Graph is best suited for analysing the relationships between different types of data.

The way it has been explained today is that the Graph tool is useful for spotting and understanding failures in a highly networked (IoT) system, for understanding customer relationships and also for spotting fraudulent activity.

Prior to the acquisition, DataStax customers were calling for tighter integration with graph, but Bosworth said that the technology proved tricky to scale initially (hence the acquisition).

We did have this problem, we had the proper back-end engine, they (Aurelius) had that modelling capability. Turns out though it was a lot harder than any of us thought to make graph scale, the way that Cassandra can scale. That’s why we had to deeply integrate it.

But it can’t be sufficient by itself. It’s almost like asking, which blade on a pair of scissors is more important? You actually need both to have a pair of scissors. In our world, what’s more important, the core DSE scalability or the graph modelling? You need them both, because the use cases are demanding those different data models.

With maturity comes challenges

Despite the improvements in capability, Bosworth was also ready to admit that there was work to do, both for DataStax and for the enterprise buyers. With the maturity of open source NoSQL technologies, and their growth in popularity, has come a race to find the best talent and a steep learning curve. He said:

They want the maturity of the application now. They want us to put more time into going with deeper integration with their existing IT infrastructure, alerting or monitoring systems, as an example. They want us to be able to automate more and more for them. If you think about the relational market this is how it played out as well.

We are past the era of asking if this technology solves the problem. We have hundreds of very public implementations at scale. Now the big problem is skills gap. I don’t have enough people that know how to do this, they are asking us to make it simpler and simpler. Because they can’t go out and hire 200 new people. You can’t speed the learning, which is the problem. You can make it easier [to use].

Even though people are being certified and trained, it’s going to be a couple of years before these systems go through real production cycles and real production. You have to just get experience, there is no shortcut. We can train people fast, you can make the product simpler, now people are getting that real world experience of learning these applications.

What next?

Whenever interviewing the CEO of one of these NoSQL companies, I always am keen to get their

opinion on how they see their companies expanding beyond the core database offering. If we retrospectively look at the evolution of the relational database companies, it’s clear that moving up the stack is inevitable. But when and how will this happen for the likes of DataStax?

Bosworth’s answer is one of the most robust I’ve heard thus far. He gave a clear indication that it will happen, but it’s going to take time and it will depend on how the market develops around companies like DataStax. For example, which companies are going to build the most successful applications using Cassandra and where will they go?

He said:

I think it will parallel the exact same way that the relational market got to that point. If you go back and study what happened, it was almost always through early partnerships. I think one would argue that one of the most seminal moments in Oracle was when SAP also started to become popular and it had Oracle as its under the cover standard.

What’s happening today is that we are seeing a lot of our start-up companies, who are building on top of DSE, is that some percentage of those will become the next great big thing. When that happens, the stack gets richer, the partner ecosystem grows. Only then the vendors can decide, how should I go up the stack? Because if you make a mistake with that, you can really damage yourself.

You have got to be really certain of where you go up the stack. We are just nowhere close to that, we have years ahead of us of continuing to make sure that the data layer stays very, very rich and robust and operationally mature. And we will embrace that partner ecosystem to build on top of us. We are more interested in making sure our development ecosystem can be satisfied, than trying to go up the stack into some unknown verticality.

And what of the relational database providers? Bosworth was not silly enough to suggest that relational technology is dead or that the likes of Oracle should be underestimated. However, he did say that those companies will find it difficult to adapt their existing business models to one that is fundamentally different to the one that is found in the open source NoSQL market.

Bosworth said:

Oracle is full of brilliant people. Oracle is a company that does $9 billion a quarter, they know what they are doing. They know how to do technology. The challenge, if I can be so bold to say this, is that historically you don’t see the old leaders also become the new leaders. A lot of the time it has as much to do with the business model, as it does the technology.

So if Oracle brought in a company that had a 1/10th cost model to customers. And suddenly you have to figure out how do you make that work inside your larger business model where I have got a 30 or 40 year history of all of our margins and how I license and how I pay reps - it’s just a harder problem to solve.

My take

Connected modern world concept © James Thew -
(© James Thew -
These companies are developing these stories well, they are becoming more interesting each time I talk to them. And from the customers speaking at the event this week, there is huge interest in what NoSQL can do for the new cloud-based applications that they desire. The message from them is that relational just doesn’t come anywhere near to cutting it for the use cases they are looking at.

That being said, this market still has plenty of maturing to do. The integrations with other players from across the enterprise stack are still taking place and the simplification of usability and management is developing at a rapid rate.

But as I hint at above, the interesting part will be when the companies focusing on higher up the application stack, using companies like DataStax to build their own technology, begin to thrive. Which companies those will be is still anyone’s guess, but that will be an interesting story to follow.

A grey colored placeholder image