Enfabrica: The Key to Scaling Compute Performance

Shrijeet Mukherjee, co-founder and, chief development officer, Rochan Sankar, co-founder and chief executive officer of Enfabrica.

Throughout history, technological innovation has enabled the greatest leaps forward in societal development. In today’s digital world, forward progress–from strengthening virtual bonds to curing diseases and advancing transportation systems to furthering automation–is propelled by a perpetually growing demand on computational power and data. Attributable to this demand, the workload on the global computing infrastructure is compounding and, if these demands are not addressed efficiently, a computing crisis could approach quickly, ultimately inhibiting societal development globally.

Distributed computing (DC) is not restrained by scalability, availability, and computing power limitations, which makes it optimal for handling enormous workloads and sustaining the growth of digitization and technological advancements. However, with the scale of computational power demands that lie ahead, computing architecture must evolve to efficiently accommodate the movement of enormous amounts of data and metadata.

Enter Enfabrica, which is developing foundational fabrics to address the major constraints of workloads in every form of a distributed system, regardless of scale, across cloud, hyperscalers, at the edge, enterprise, 5G and 6G networks, automotive infrastructure and more.
"Our team, which includes individuals with the perspective and background of creating successful products, has developed a solution with universal appeal and value," says Shrijeet Mukherjee, cofounder and chief development officer of Enfabrica.

It is essential for computing components inside storage and compute elements to efficiently communicate and exchange massive amounts of data with each other. Enfabrica identified that the key to scaling performance and capacity of DC rests on scaling and optimizing the input/output (I/O) connectivity infrastructure within servers, data centers, and compute clusters. So, Enfabrica has developed groundbreaking hardware, software, and system technologies that enhance I/O architecture to better support the rapidly expanding DC and its workloads.

"Enfabrica is seeking to build market-leading foundational fabrics that can be used by any customer who seeks to build computing infrastructure," says Rochan Sankar, co-founder and chief executive officer of Enfabrica.

Only a small percentage of products are truly disruptive in the sense that they cause a revolution in an industry or change the world, but Enfabrica's hardware, software, and system technologies qualify. Its product builders, architects, developers, and ethical hackers have created foundational fabrics for DC, and its substrates glue compute, storage, and networking in a harmonious and scalable fashion, enabling faster communication between them.
In addition, Enfabrica has joined a market of burgeoning opportunity, as a recent report from 650 Group predicts the interconnect market (the technology that enables communication between data center infrastructure, systems, and semiconductors) to grow to $25 billion in revenue by 2027.

Solving the I/O bottleneck significantly increases the speed with which data is transferred and processed within data centers, between compute and storage, and various computing components like CPUs, GPUs, purpose-built accelerators, or field-programmable gate arrays. Enfabrica's solutions will transform the IT infrastructure so workloads can be efficiently mapped and scaled across heterogeneous systems and components in a way that communication between them is not hindered.

Our team, which includes individuals with the perspective and background of creating successful products, has developed a solution with universal appeal and enduring value

Simply, today’s server I/O architecture cannot handle the growing compute demand being fed by digitalization and an increasing dependency on technologies such as artificial intelligence and machine learning. Enfabrica's refactored combination of silicon, system hardware, and software, though, will optimize the server I/O architecture to be more performant and efficient. The adaptability of its solutions can tackle the challenges of perpetually increasing workload demand by enhancing the performance of distributed systems while being scalable for architectural change in the computing industry.