Startup Lucata Corp. today announced that it has raised $11.9 million in funding to expand industry adoption of its Pathfinder hardware platform, which speeds up graph databases by allowing their code to jump between processors.

Graph databases are a type of database that can store not only records such as sales transactions but also contextual information on how those records are connected with one another. That information can be useful in many situations. Enterprises rely on graph databases to perform a variety of tasks ranging from finding ways of optimizing their supply chains to generating shopping recommendations for customers. 

The faster a graph database can analyze data, the faster a company can obtain the insights it needs to carry out an action such as displaying a personalized shopping suggestion to an online buyer. Lucata says its Pathfinder hardware system (pictured) improves processing speeds by “orders of magnitude” in some cases. The system also lends itself to accelerating artificial intelligence training, the startup claims.

Lucata, which was founded by three computer science and engineering professors, says Pathfinder provides the speedup using a patented innovation dubbed migrating threat technology.

Large-scale graph database environments usually run on not one but multiple servers. Consequently, it’s often the case that a piece of data stored on one server is processed by code running on a different server. This normally requires moving the data being processed to the server running the code, which slows down computations.

Lucata’s solution is to effectively reverse the process: its migrating threat technology moves the code to the data. This is accomplished by temporarily transferring the processor thread that contains the code from its original host server to a chip inside the server containing the information being processed. Historically, the startup says, it was possible to move threads between processors located in the same server but there were few ways to move threads across different servers, as Pathfinder does to optimize graph databases.

Each Pathfinder system has at least 8 terabytes of flash storage and 512 gigabytes of memory. More than 1,000 server racks containing eight Pathfinder systems each may be linked together to form an infrastructure cluster for running graph databases. According to Lucata, a graph database environment powered by Pathfinder can be equipped with more than 4 petabytes of memory and 64 petabytes of speedy flash storage.

Lucata says the increased processing speeds facilitated by its system enable enterprises to run more sophisticated data analyses using their graph databases than they could before. Furthermore, the startup says, its technology simplifies developers’ work because they have to spend less time optimizing tasks related to moving data within the hardware cluster. 

“With the increasing need to find deeper insights in ever-growing sparse datasets, there is a compelling need for the patented technology Lucata has developed,” said Lucata Chairman and Chief Executive Officer Michael Mallick. “Commercializing this groundbreaking technology has required years of development effort. We are now at an inflection point when the technology evolution is aligned with market needs.”

Lucata will use the newly announced $11.9 million round to step up go-to-market activities. The capital was provided by an investor consortium that included Middleburg Capital Development, Blu Ventures Inc., Hunt Holdings and others. 

Image: Lucata