Earlier this week, Pivotal made headlines when it released a commercially supported version of the open-source Cloud Foundry platform-as-a-service (or PaaS) for building and running applications. In the same breath, the company announced supported services to store and analyze large quantities of data.
Vendors don’t typically launch major cloud-computing and big data products on the same day; it’s usually one area or the other. But Pivotal executives see data analytics as a valuable resource for running applications, and they think querying data should be just a few steps away from the management of a cloud, not in a whole other universe. And that mindset helps Pivotal stand out in the PaaS market as well as the data market.
To learn more about the merger of cloud tech and data science, we talked to Todd Paoletti, Pivotal’s vice president of product marketing. He discussed how far along data science is inside enterprises, how easy it is to create a repository of data that you can process, and how developers can use analytics to turn out better applications.
Here’s an edited transcript of our conversation.
VentureBeat: What made you guys want to roll out big data services plus commercially supported PaaS together? Why are they coupled like this?
Todd Paoletti: Well, the PaaS is extremely valuable in its own right, because Pivotal still is very much about choice, being able to run your application, your environment, your workloads in the cloud of your choice, being able to leverage the development environments of your choice, etc. So Pivotal CF is extremely valuable in its own right for those purposes — collapsing the time to value for deploying applications on a variety of clouds or systems.
But the Pivotal One data services are accessible and integrated within that environment. We are getting customers closer to the time-to-value notion, effectively within the construct of Pivotal CF. A handful of clicks will enable a user to set up a big data cluster. A few more clicks will enable the same customer to begin running analytics and leveraging data in that cluster or analytics that are overseeing applications that are stood up on that cluster, to get value out of it. We see those services as really important value-adds for the dev-ops team in the enterprise, but these are also really important from the spirit of helping the enterprise get to value more rapidly through services that are right there on the screen for them.
VentureBeat: Is it really possible to set up a Hadoop cluster with a few clicks?
Paoletti: It is, I would say, in spirit, a few clicks. In a demo, you can see a Pivotal interface. Presented in the interface are service tiles that represent these Pivotal One data services as optional service tiles, and from there the installation and deployment of those services essentially act as any other application that’s being installed and deployed by the PaaS.
VentureBeat: How easy is it from there?
Paoletti: Well, that’s the value of the PaaS in and of itself. I mean, effectively, we are creating an environment, a framework, for a dev-ops team or an operations team to manage installations, manage workloads, provision applications, deploy, scale up, and scale down entirely through a controlled environment, and we take what could take weeks down to days or hours, depending on the application.
Editor’s note: Our upcoming DataBeat/Data Science Summit, Dec. 4-Dec. 5 in Redwood City, will focus on the most compelling opportunities for businesses in the area of big data analytics and data science. Register today!
VentureBeat: Are companies actually showing interest in paying for both the PaaS and the data services?
Paoletti: Three days ago, it was the first time we actually had available or announced a commercial version of the on-premise PaaS system. Time will tell if we got it right. I will say that our customers have been asking for this kind of capability for a while, so we think we’ve got it right.
VentureBeat: What if companies can’t think of good questions to ask even if they want to analyze data?
Paoletti: Well, then we wouldn’t recommend that they buy the service, honestly. It doesn’t help us if someone doesn’t get use out and value out of a service of the gate. If the customer knows what they want out of big data, they’re going to be down a maturity curve with the software we’re providing them. We’ll absolutely help in and of itself if they are in a phase where they are trying to get value of big data. That’s where our elite-level data science labs organization comes into play.
That team is one of the most sought-after teams within the company, and effectively, they go into an organization that says, “We know we need to get value out of big data,” or “We’re capturing it, but we’re trying to understand how to convert that valuable data into insight and analysis.”
That’s what the team is designed to do, and that’s what they do very, very effectively.
VentureBeat: Do you find that companies already employ data scientists?
Paoletti: The notion of data science, if you look at the history of that world, is relatively new, and so there are not that many data scientists out there. There are not enough data scientists to go around, which is why our team is in high demand. Not every enterprise has a data science team. They are more times than not small. But they are growing.
VentureBeat: Are you seeing that developers of the applications that can run on the PaaS want to ask data science questions?
Paoletti: We are, and I think it’s a combination of algorithmic data and application development — what we call data-driven applications. It’s what Google does really, really well and what Facebook does really, really well. They see algorithmic updates in real-time data, so that the app behaves with input from big data. One use case is to help protect against fraud. Another use case is making purchase recommendations through pattern recognition on buying patterns. That’s what Amazon does really well.
VentureBeat: How confident are you that other PaaS sellers will stick their own data services on top of PaaSes, instead of separate from or next to them?
Paoletti: If they do it right, they will, in part because we’re seeing that customers want these things together.
VentureBeat: What will Pivotal come up with next on the data side?
Paoletti: Well, we’re going to continue to invest in advanced systems and tools that run across our overall stack, and we will continue to invest in integrating those components of our data stack together, so that companies can use everything more effectively.
From massively parallel database technology to Hadoop subsystems to predictive analytics, and Hadoop for the purpose of real-time applications — those advances we’ve made already. You can imagine we’re going to continue to invest in the most cost-effective, flexible, and advanced data architecture to help support the rapid development of big data apps and big data analytics.
VentureBeat: Which do you think are more in demand: data services like those introduced this week, or the commercially supported Pivotal CF?
Paoletti: I think the trend is that this stuff is going to be really, really important. I can’t give you a concrete answer. We’ll see how the uptick goes. What I will say is — look at the enterprise adoption of cloud services, infrastructure services, platform services — like an IDC report or a Gartner report. There’s massive acceleration in that space. Cloud Foundry and Pivotal CF will participate in those growth cycles.
Separately, if you were to look at an IDC or Gartner report, our data services are going to participate and hopefully exceed growth from those markets independently.
So the big growth in cloud utilization and cloud platform adoption absolutely will be big growth for Pivotal CF. Big growth in big data will be big growth for Pivotal One. We believe the combination and the intersection of those two will help us accelerate both even faster than those markets are accelerating, if that makes sense.
To be pithy about it, you can eat peanut butter without chocolate, and you can eat chocolate without peanut butter. There’s a market for them that are discrete, but some people would argue they’re better together.
But the value of the PaaS system improves with the number of data clusters that are accessed by it and managed by it. The ease of managing applications and managing Hadoop clusters on top of that through the PaaS will perpetuate the adoption of Hadoop in the enterprise. So we think that they are symbiotic, but they do not necessarily have to be sold together or consumed together.