Industry Collaboration Drives 100G Foundation for Supercomputer Infrastructure

Supercomputers were first introduced in the 1960s, and were considered to be at the frontline of process capacity and calculation speed. While the supercomputers of the 1970s used only a few processors, in the 1990s machines with thousands of processors began to appear. By the end of the 20th century, massively parallel supercomputers with tens of thousands of "off-the-shelf" processors were being used.

For several decades now, these supercomputers have played a critical role in the industry, opening the door for global innovations in science, engineering and business. Today, at the rate data streams are increasing, research facilities and data centers are realizing the need for a fast and power-efficient network is a necessity to compete. We live in a world where we want to have answers at our fingertips, which means that everyone needs to have information processed faster.

There have been some recent developments that herald advances for supercomputers in architectures, algorithms, software and applications, especially as the supercomputing industry moves towards the ExaScale challenge of building a supercomputer that can deliver 1 million, million, million calculations per second by 2020.

Singapore’s Agency for Science, Technology and Research, known as A*STAR (News - Alert), Computational Research Centre and Obsidian Strategics have together  an expert team of optical network and technology partners to rewrite history on distributing computational work over globally dispersed sites. This collaboration will result in what is being referred to as a galaxy of supercomputers designed to open the door for global innovations in science, engineering and business.

Obsidian Strategics is the developer of Longbow and Crossbow, which is a series of InfiniBand products featuring range extension, routing and encryption. Longbow technology allows an InfiniBand fabric, normally a short-range network used in supercomputers and data centers, to be securely extended via optical fiber over global distances. Crossbow technology provides very high performance routing of InfiniBand traffic between many separate subnets, enabling unprecedented scalability and manageability.

InfiniBand is next in line to turn the tide for supercomputers. The galaxy of supercomputers will enable the aggregation of many globally distributed supercomputers into a hive-mind of enormous scale. The purpose of this collaborative initiative is to demonstrate the feasibility of the 100G infrastructure and how it will lead to the future of global supercomputer developments in architecture and scheduling algorithms.

Dr. David Southwell, co-founder and chief visionary officer at Obsidian Strategics, had the following to say, "Supercomputers have played crucial roles in military, higher-education, R&D labs and industry for decades. Today, at the rate data usage is increasing, the uses for high performance computing have also increased.  Most supercomputers today are built around a very capable open-standard LAN technology called InfiniBand, but this is limited to very short reach connections.  Our work on mission critical global communication requirements from the U.S. government has allowed us to pioneer the technology capable of transparently extending InfiniBand over arbitrary distances. The result is a highly deployable communications fabric suitable for demanding applications in not only scientific computing, but also storage, data center, cloud and enterprise environments. We couldn't be more pleased to be a part of this groundbreaking collaboration."

Listed below are a few concepts of what we can expect to see;

• InfiniCortex, galaxy of supercomputers– A new architectural approach to meet the challenge of building the next generation of supercomputers by enabling the aggregation of many globally distributed supercomputers into a hive-mind of enormous scale. This pilot program is an example of how research centers, such as A*STAR CRC, can meet the challenge of ExaScale computation, in the same spirit of international cooperation and burden sharing as seen with LHC, ITER and SKA.
• Enabling long-distance connection– Initially developed to meet the needs of military and intelligence community networks, Obsidian's Longbow technology provides transparent range extension of local area InfiniBand traffic over standard WANs such that remote supercomputer fabrics can merge across global distances. High-grade encryption and authentication functions enable secure international sharing of highly valuable and critical supercomputer resources.
• Demonstrating feasibility and lasting value –To demonstrate the feasibility and lasting value of the pilot platform, various workloads will exercise the InfiniCortex at SC14, including real-time analysis and control of a plasma physics experiment at the Tokyo Institute of Technology involving distributed data processing concurrently occurring in Singapore, Georgia Tech and equipment at SC14 exhibition floor.

The team will demonstrate transmission of data at a speed of 100 Gbps across the Pacific via subsea optical cables to the Supercomputing 2014 (SC14) conference. We can expect to see this record breaking demonstration deliver data ten times faster than previously recorded transmission speeds between Asia and North America.

A*STAR CRC’s senior director, Dr. Marek Michalewicz, made the following comments "As our teams began the extensive process of developing the InfiniCortex initiative, we realized that we would require input from a variety of global resources. We are proud to be able to provide the necessary mathematical tools and related software to develop a geographically distributed supercomputer of supercomputers based on this global collaboration between HPC clusters. This demonstration is the culmination of the extensive efforts of all those involved and we are looking forward to unveiling the results at SC14."