Data warehouse startup Yellowbrick Data Inc. is emerging from stealth mode today after four years in the making, backed by $44 million in venture capital funding.

The company is touting what it says is a “revolutionary” data analytics platform for hybrid cloud environments that’s designed for “direct flash queries” with results returned in a matter of seconds or minutes. It claims to deliver results from queries up to 140 times faster than existing data warehouses can do, while taking up 30 times less space in the data center.

Those impressive claims are backed by some equally notable investors, with big-gun venture capital firms such as Draper Fisher Jurvetson, Google Ventures, Menlo Ventures, Samsung Ventures and Third Point Ventures all participating in the Series A funding round.

Founded in 2014, Yellowbrick Data built its platform, which includes both software and a hardware appliance, after coming to the conclusion that data warehousing needed a major overhaul to keep up with the endless flows of data organizations are dealing with today.

The company said in its pitch that current data warehouse technologies are extremely bloated and a huge drain on resources. Moreover, enterprises are struggling with fragmented ecosystems that spew out data from multiple sources, which will only become more complex as emerging technologies such as the “internet of things” and artificial intelligence become more commonplace. Yellowbrick said it’s bringing back “simplicity” to data warehouses, which are central repositories of integrated data from one or more disparate sources used to provide business intelligence.

There are two components to Yellowbrick’s platform. The first is a single 12-inch hardware appliance that can be swapped in for six seven-foot hard disk-based data storage cabinets, making it 97 percent smaller than existing systems. The appliance is 100 percent flash memory-based, meaning it can perform queries on data at the same speed are regular random access memory. The reliance on flash also means massive savings on data center cooling and energy costs, the company said.

Wikibon analyst James Kobielus told SiliconANGLE it’s interesting that Yellowbrick decided to build an appliance at a time when most new data warehousing solutions are instead cloud-based. Appliances were in vogue 10 years ago but have been largely eclipsed by cloud-based offerings in terms of customer adoption over the past several years, Kobielus noted.

However Neil Carson, Yellowbrick’s co-founder and chief executive officer, told SiliconANGLE the company is a strong believer in “hybrid cloud analytics,” hence the idea is to build and deliver products across that spectrum, including an appliance.

“This includes edge and the data reduction of things, as well as on-premises for companies that like or legally have to work that way,” Carson said. He added that hybrid cloud analytics also includes colocation for companies that need to operate their own infrastructure and run this alongside public clouds for economic reasons, as well as public clouds.

The other component to Yellowbrick’s platform is the data warehouse software itself, which the company said has been architected specifically for native flash-memory queries.

The software combines networking, storage, CPU, scheduling and an optimized analytic database with autonomous administration, and is designed for high levels of data ingest and processing. This enables users to run “mixed workloads, including ad hoc queries, large batch queries, business reports, ETL processes and OBDC inserts simultaneously,” the company said.

Nonetheless, Kobielus said he was skeptical of Yellowbricks’ claims around the performance of its data warehouse, noting that it hasn’t published any benchmark tests supporting its assertions around its speed and savings.

“In-memory databases are the mainstream now, so at first glance this doesn’t seem to be anything differentiated,” Kobielus said. “Their ‘native flash queries’ does not seem all that great a differentiator, in terms of query performance, vis-a-vis MemSQL’s in-memory architecture, which can move data to a flash SSD-based column store for fast, large-scale analytics in 10s of terabytes in memory and several petabytes on disk.”

But Carson said that benchmark tests can be easy to rig, and that Yellowbrick preferred to come out of stealth backed by “referenceable customers” that could back up its performance claims. The company’s list of early adopters is certainly impressive, and includes the telecommunications firm TEOCO Corp. and online retailer Overstock.com, Inc., among others. Carson said that every number Yellowbrick has mentioned regarding its performance was logged by its customers.

And those customers have given Yellowbrick’s platform a resounding thumbs up. “After one week of testing, it was clear we were looking at something groundbreaking,” TEOCO Chairman and Chief Executive Atul Jain said. “Yellowbrick will be our platform going forward. We’re replacing 24 racks of our prior solution with a half of a rack of Yellowbrick appliances.”

Yellowbrick said its data warehouse appliance and platform have been generally available since September 2017, and have already seen wide adoption by customers in the retail, telecommunications, logistics, healthcare and other industries. With regards to the latest funding round, Carson said the plan is to grow the company’s sales and marketing teams in order to expand its reach, and also expand its data warehouse into new environments.

“We’re planning to announce our public cloud offering in the next 18 months,” Carson said.

Image: Pete Ashton/Flickr