Splice Machine: A Realtime SQL Database on Hadoop

Mark Boyd
Oct. 14 2013, 10:00AM EDT

Splice Machine prides itself on being the only transactional SQL-on-Hadoop database aimed at giving developers an easy-to-learn tool for managing big data in the enterprise. ProgrammableWeb spoke with Splice Machine's CEO Monte Zweben at the recent API World and Data Week conference in San Francisco - where they won a Top Innovator award - to discuss how API developers can access the best-of-both-worlds SQL and NoSQL database engine to return contextual and personalization data in realtime.

"Developers have had to make a big leap away from relational databases to get the scale they need. Which is why so many have moved to MongoDB," Zweben said, obviously hoping that Splice will have a similar trajectory to the New York database software company, now rated as most valuable internet startup in NYC, having raised a further $150 million at the start of the month to take the company to a value of $1.2 billion.

MongoDB has been able to build a strong foothold against relational database competitor and market leader Oracle with both a lower cost product and a more scalable approach to storing big data.

Splice Machine hopes to sit somewhere between Oracle's relational database services and MongoDB's No SQL approach.

"There so many advantages to using a SQL database, if you can provide developers with a scalable platform," Zweben said.

"Our main message to the web developer community is that you can get the advantage of the NoSQL databases and still get all of the SQL services that backend developers know and love.

"For example, if you need to build a large scale e-business application, or need scalability and need to transact very quickly with fast reads and writes, our database helps do that, and if you need analysis: our tech can do both."

Zweben is quick to let the PW audience know that Splice Machine can be accessed via API. They also offer built-in connectors to provide turnkey integration with Tableau Software, for example. "At the core of the Splice Engine are connectors that allow developers to access big data via APIs. We are built on top of Hadoop stack, with a RESTful API."

This enables developers to build Splice Machine into their data architecture in order to create more meaningful end user experiences drawing on both proprietary customer data (leading to better personalized communications) and demographic and locational datasets (providing greater contextual responses).

"One reason why we built our engine is to support personalization services. For example, we have a client doing personalized direct marketing services, another customer is personalizing emails. Suppose you want a dynamic emailing content service that opens up different content depending on the device or what the weather is like, that email can have a RESTful call that can do that.

"Many of our customers are trying to blend data sources. They may have a large history of data stored in our database, they may have an application with realtime data attributes they want to utilize in their business logic, and they may have acquired data services and use a relational database. They want data results in realtime, in milliseconds."

Zweben believes Splice takes a unique position in the market and lets developers start building data workflows quickly into their enterprise and end-user products.

Aside from the Top Innovator award at DataWeek, Splice Machine has already been turning heads amongst industry pundits who believe the technology may offer a third way between Oracle and MongoDB. Zweben explained the company's unique market position:"Our relational database is so scalable: we are the only realtime SQL database built on Hadoop."

Mark Boyd is a ProgrammableWeb writer covering breaking news, API business strategies and models, open data, and smart cities. I can be contacted via email, on Twitter, or on Google+.

Comments