Earlier this month, startup rENIAC announced it has received Series A funding led by Intel Capital with participation from UMC Capital, Leawood Venture Capital, Divergent Capital, Bullpen Capital, Upheaval Investments, and J-Angels. The funding will be used to build a sales and marketing team and ramp up customer adoption, according to Prasanna Sundararajan, CEO and co-founder of rENIAC.
A drop-in addition to database clusters that requires no changes to software or data architecture, the rENIAC Data Engine (rDE) improves the performance of latency-sensitive, data-centric applications, and is targeted at organizations that need a database accelerator to achieve lower latency for consistent SLAs, increases in throughput, reduced TCO.
The engine is comprised of storage, network, and compute engines, and leverages FPGA, CPU, and tiered storage (DRAM+SSD) to accelerate data workloads. rDE supports acceleration for SQL, NoSQL, and graph databases and can be implemented as an Apache Cassandra I/O accelerator; distributed, persistent cache; or distributed, scalable storage engine for graph and relational databases.
The rENIAC Data Engine (rDE) is a drop-in addition to database clusters that requires no changes to software or data architecture.
With rENIAC already in use by organizations including a major telco and major movie studio, said Sundararajan, the data engine is aimed at powering the increasing requirements of data center and edge applications, especially those based on AI and machine learning. In particular, rDE is useful for fraud prevention, payment processing, identity resolution, and recommendation engines. It can be deployed on-prem, hybrid, or in the cloud and—by combining the power of FPGAs with proprietary software—helps customers reach their goals of reduced CPU usage and node count, increased performance, and consistent, predictable SLAs that the company says is unattainable through a traditional database alone.
Driving Customer Need
According to Sundararajan, rENIAC is in active discussions with companies in a number of verticals but most customers generally fall into two buckets. One bucket is comprised of customers that have a Cassandra pain point and the other bucket is made of customers that don't necessarily have Cassandra, but do have an immense problem with the volume of data and and can't get enough out of their existing infrastructure.
As far as rENIAC's focus on the open source database world, Sundararajan noted that companies such as Google or Facebook undoubtedly have groups of software and systems engineers and DevOps people to optimize their infrastructure so they meet their business needs and performance goals, but that is not true for most enterprises. While proprietary software vendors are prepared to support customers with their own products and services, open source software may present a different set of problems for organizations faced with the need to scale. "You can only get so much throughput and density per node and, before you know it, if you have a large dataset, you have hundreds if not thousands of nodes, and you need an army of engineers to manage the complexity."
Most rENIAC customers fall into one of two buckets now: customers that have a Cassandra pain point and customers that may not necessarily have Cassandra, but do have an immense problem with the volume of data and can't get enough out of their existing infrastructure.
The choice of Cassandra as an initial focus is because "Cassandra usage is here to stay and growing," said Sundararajan. It is used in mission-critical applications where business KPIs are associated with it. Using Cassandra as a starting point, rENIAC's roadmap includes plans to expand support for other open source database platforms such as JanusGraph as well as MySQL and MariaDB.
Next Steps
The next steps for the company are to scale and increase market adoption, said Sundararajan. "The vision for the company is to help companies manage the growth of data without sacrificing performance and without having to change their software infrastructure. You want to make it as friction-less as possible."
According to Sundararajan, the amount of data needed for real-time, customer-facing applications is impossible to operationalize when managed through software alone.
The drop-in data accelerator enables customers using open source databases to "right size" their existing data architecture while ultimately improving the end user experience.