Follow Datanami:
September 30, 2020

TigerGraph Offers Free Graph Database for On-Prem Analysis

(ra2 studio/Shutterstock)

TigerGraph today rolled out a new deal that allows customers to store up to 50 GB of data in a distributed graph database running on-premise, matching what it already offered in the cloud. The company also welcomed more than 3,500 attendees to its inaugural  Graph + AI World conference, which included keynotes from customers like Jaguar Land Rover and UnitedHealth.

Banks and healthcare companies have some of the most compelling use cases for graph analytics, including anti-money laundering (AML) and drug discovery. However, these companies are also among the least able to take advantage of cloud-based graph offerings, such as TigerGraph Cloud, due to stringent data regulations.

That’s the reasoning that went behind TigerGraph’s announcement today to give away copies of TigerGraph Enterprise Edition, its full-featured graph database. The software is limited to storing up 50 GB of data, which is enough for prospects to see if the solution will work for them, says Todd Blaschka, TigerGraph’s COO.

“There are some customers that cannot move their data to the cloud due to sensitivity restrictions,” Blaschka says. “Developers have been using [a free version] that is designed for a single machine, but they’ve been saying to us, hey we want what is also available on TigerGraph Cloud. Can you make it available?”

TigerGraph unveiled its cloud offering two years ago, and made it generally available last year. When it launched on AWS a year ago, the SaaS offering was free for databases running on up to 4 CPU and encompassing 8 GB. The cloud offering has been very popular, and today has more than 2,000 users, Blaschka says.

The cloud offering is the leading vehicle for getting TigerGraph into the hands of prospective customers, says Gaurav Desphande, vice president of marketing for TigerGraph.

Graph databases make sense of connected data (Sergey Nivens/Shutterstock)

“If you’re a cloud customer, no problem. Come to TigerGraph Cloud, and use the free tier to build applications,” Depshande says. “We provide an open source graph algorithm library. We provide graph studio, which is the best data science GUI for graph databases for connected data analyses. And then we have partners providing additional applications and starter kits that provide you with the baseline schema, a set of services, and the workflow–all out of the box, all at our cost.”

TigerGraph doesn’t charge for compute or storage in the cloud offering, and now, with the free version of TigerGraph Enterprise, it’s giving customers the same opportunity on-prem (as long as you supply the compute and storage). TigerGraph is there to provide 24/7 support for those who put large-scale graph applications into production. “You can build applications in the cloud or on premise,” Desphande says. “If you see the value after you get started, please pay us something–we’re a tiny startup company. We need to pay our bills also.”

TigerGraph was founded several years ago by Yu Xu, a distributed systems expert who previously worked at Teradata and Twitter, to build what he terms a “third generation” graph database that’s designed from the get-go to run in a distributed manner. Even with only 50 GB in the free version, customers can take advantage of TigerGraph’s distributed nature to process the data in a cluster, bolstering performance.

The free version of TigerGraph Enterprise isn’t limited to development or test use cases, and can be used in a production setting. Many problems can be tackled with 50 GB, which is the equivalent of 150 GB on competing graph databases due to the way they tend to “blow up” data, Deshpande says. The database is capable of scaling up to support hundreds of terabytes of data across a cluster.

In addition to running as an operational database, TigerGraph can be used by data scientists to model data using built-in graph algorithms. Those algorithms can very quickly thanks to a partnership the Redwood City, California company recently formed with Xilinx, a provider of field programmable gate arrays (FPGAs).

“Hardware-accelerated machine learning is another top use case where people are combining TigerGraph with other hardware like Xiliinx’s FPGA to accelerate community detection” of money laundering and fraud rings, Deshpande says.

This week also marked the company’s first foray into virtual conferences. TigerGraph is the host of Graph + AI World and its main sponsor, but the idea is to build the conference into a place for the graph database and AI communities to come together to share ideas.

Among the speakers planned for the event are Edward Sverdlin, a vice president at UnitedHealth Group, which used TigerGraph to analyze medical codes for potential COVID-19 patients; and Harry Powell, the director of data and analytics at Jaguar Land Rover, which used TigerGraph to accelerate supply chain planning operations that were disrupted due to the pandemic.

So many people signed up for Graph + AI World that TigerGraph was forced to add an extra day. All told, the event featured one day or tutorials and two days of business and technical keynotes. The event is also open to TigerGraph’s competitors, Blaschka says.

“We would love to have more graph companies participate and share what they’re doing,” he tells Datanami. “We embrace it. We want to drive the industry forward as a whole, just like Databricks is doing with Spark + AI Summit. We’re just the ones organizing it.”

Related Items:

TigerGraph Cloud Goes Live as Firm Raises $32M

TigerGraph Launches Cloud Database on AWS

TigerGraph Emerges with Native Parallel Graph Database

Editor’s Note: This story has been corrected. When TigerGraph launched its free cloud offering in 2019, the service was available for databases running on up to 4 CPU and encompassing 8GB of data, not 750 GB.  Datanami regrets the error. 

Datanami