A3Cube Launches, Promising Faster Networking for HPC and Big Data

a3cube

The home page of networking startup A3Cube.

Silicon Valley startup A3Cube emerged from stealth mode today, and is looking to introduce a dramatic shift in network architecture. The company is backed by five years of research and development, and has been in business for a little over a year.

A3Cube has announced a "brain inspired" PCI Express Network Interface Card (NIC) designed to eliminate the I/O performance gap between CPU power and data access performance for data center, HPC and big data applications.

Nano-Second Latency

A3Cube says its RONNIEE Express platform elevates PCI Express from a simple interconnect to an intelligent network fabric that solves performance bottlenecks inherent in PCIe. A3CUBE’s In-Memory Network technology allows direct shared global memory across the entire network. This data plane enables storage that combines supercomputing’s massively parallel operational concepts and an I/O interface eliminating central switching and network overhead.

The company says this In-Memory Network discards the protocol stack bottleneck and replaces it with a direct memory-to-memory mapped socket, producing disruptive performance enhancements while using commodity hardware.

“A3CUBE’s In-memory Network Fabric leverages an innovative approach to transforming HPC, Big Data and data center environments in order to drive greater performance and efficiencies in the network and storage systems," said Bob Laliberte, senior analyst at ESG. "A3CUBE is extending PCIe capabilities in order to deliver a next generation network that it claims will overcome traditional network bottlenecks utilizing a high performance (Nano-second latency) and massively scalable architecture.”

Touts "Radical" New Design Approach

“Today’s data center architectures were never designed to handle the extreme I/O and data access demands of HPC, Hadoop and other Big Data applications,” said Emilio Billi, founder and CTO of A3CUBE. “The scalability and performance limitations inherent in current network designs are too severe to be rescued by incremental enhancements. The only way to accommodate the next generation of high performance data applications is with a radical new design that delivers disruptive performance gains to eradicate the network bottlenecks and unlock true application potential.”

RONNIEE 2S is a PCIe-based intelligent NIC designed to maximize application performance using a combination of hardware and software. It provides multiple channels with sub 1 microsecond fast direct remote I/O connections. RONNIEE RIO is a general purpose NIC supporting Ethernet and memory-to-memory transactions in a 3D torus topology that can plug in any server equipped with a PCIe slot. This empowers a scalable interconnection fabric based on a patent-pending shared memory architecture that implements the concept of distributed non-transparent bridging.

RONNIEE 3 is a card designed to extend the capability of RONNIEE 2S and optimized for high performance data environments. The In Memory Network provides full support for memory-to-memory transactions without the usual software overhead.