Boaz Raufman JethroData Co-founder and CTO
We got great help from Simba...how to set the ODBC attributes, which ones were relevant...help for tuning and configuration. Every issue we encountered, we got answers very, very quickly.
Building the perfect ODBC driver
Some years ago, JethroData’s three founders—Boaz Raufman, Ronen Ovadya, and Eli Singer—envisioned a BI-enhancing database that could index Big Data, accelerate query performance, and scale to support distributed petabyte-size data clusters. They led the development of the company’s namesake database technology: a unique engine for Big Data that combines search-engine indexing technology with modern column store database design into a single solution. Jethro customers can achieve lightning-fast query performance, even when their data is massively massive (and growing…massively).
In developing its signature database acceleration software, JethroData recognized a need to provide 1) ODBC connectivity to multiple data sources, and 2) SQL access to data, even when that data was stored in a NoSQL data store.
“JethroData technology can run on many storage types,” explains JethroData co-founder and CTO Raufman. “We have a unique architectural separation of everything, from compute layer to storage layer. Since we were developing a database, we knew from Day One we would have to provide integrative components [including standards-based ODBC]. If we have problems with the connectivity, all the work we’ve done underneath isn’t worth anything.”
As the JethroData team assessed the best way to deliver ODBC connectivity as part of its database-acceleration technology, a meeting at a Hadoop event proved fruitful. Company leaders were introduced to Simba stakeholders, and learned about Simba’s ODBC solutions, including drivers and development environments.
“We needed to do ODBC,” begins Raufman, “so we considered two options: either use a tool with infrastructure like Simba, or try to develop it ourselves. We recognized that ODBC is complicated, that it is not our core business. We are database developers.”
The choice: off-the-shelf or code with SimbaEngine?
One challenge: Jethro’s proprietary core client-server protocol need didn’t lend itself to the convenient, “snap-in” connectivity a commercial driver might provide. And commercial drivers are by nature more limited in scope of supported data sources.
“We are developing the full stack here from the driver down,” says Raufaman. “We knew we had to have our own driver. We cannot just take something from the shelf.” Jethro started evaluating the SimbaEngine Software Development Kit.
The option to build a custom ODBC driver with the SimbaEngine SDK offered Jethro flexibility, convenience, connectivity to multiple BI tools (like Tableau and Qlik), comprehensive support, and better SQL-query coverage than Jethro could achieve with alternative approaches (like extending an open-source driver).
With collaborative support from the Simba Technologies’ engineering team, Jethro developers began coding an ODBC driver using the SimbaEngine SDK. Raufman cited the ease with which his team was able to extend SimbaEngine SDK sample code into a practical prototype. He also noted the support Simba provided along the way.
“We got great help from Simba,” explains Raufman, “how to set the ODBC attributes, which ones were relevant…help for tuning and configuration. Every issue we encountered, we got answers very, very quickly.”
What’s next for JethroData connectivity? “Deeper integration with existing BI tools and broadening the integration,” answers Raufman, concluding, “Our use case is connecting to BI tools. Every JethroData customer will have the [SimbaEngine SDK-built] ODBC component as part of the solution.”
(Download the Jethro Data case study in PDF form Here.)