By Alastair Hartrup, Global CEO of Network Critical.
The speed and processing power of modern day computers is astonishing. High Performance Computing (HPC) technologies are developing at an astronomical rate.
The United States and China are locked in a computer processing race much like the 1960's race to the moon. According to the National Security Agency (NSA) and the Department of Energy (DOE), China is winning.
Today, the competition for HPC superiority is equally as important as the 1960's race for space dominance. China's Sunway TaihuLight System is a High Speed computing system that is capable of 93 PetaFLOPS. That is 93 quadrillion Floating Point Operations per second. The Sunway TaihuLight is ranked #1 on the TOP500 list of super computers.
What is a Floating Point Operation FLOP? For the purposes of this discussion, we can agree that a Floating Point Operation is a calculation performed by a High Performance Computer. Note that it is not a bit, but a complicated mathematical computation out to multiple decimal points. HPC processing power is described in terms of FLOPS, MegaFLOPS, GigiFLOPS, PetaFLOPs and so on. A PetaFLOP is one thousand million million FLOPS. So, now we understand the power of HPC. The calculating speed and power is mind boggling.
According to Computerworld, supercomputing dominance provides an enormous economic engine. It provides the firepower for modern achievement: performing cutting edge research, developing new medicine, building better products, improving infrastructure, and designing innovative weapons systems. Regardless of which country is number one, all countries in the race reap benefits.
Yet supercomputers are only one avenue towards profiting from HPC. If clusters of smaller computers are networked to provide massive computing power, the price tag is much more attractive than for a single monolithic supercomputer. According to Market Research Future, the global HPC market is forecast to grow to £25 Billion by 2022 at a CAGR of 5%.
The two architectures for HPC, the supercomputer and clusters of smaller computers, both share the same Achilles heel: network security. Centralised super computers are expensive resources, and so computer clusters are used instead. However, clusters are networks and therefore vulnerable to internal and external threats. As such, network security is vital to HPC advancement. This is where intelligent TAPs and Packet Brokers come in handy. TAPS and Packet Brokers monitor and map network traffic through connected appliances, ensuring complete network visibility.
Valuable computing assets are sure to be a prime target for hackers. Simultaneously, HPC and Artificial Intelligence (AI) technologies are being used to protect computing resources. As security appliances are designed to learn the normal traffic patterns of an organisation, and then to quickly detect and block anomalies, high speed networks will become less vulnerable to and will recover more quickly from attacks. Such security advancements brought about by HPC and AI will benefit all networks.
China has the #1 super computer, while the United States and Europe are currently the largest HPC markets. The awesome computational power HPC technology will certainly bring forth many advances in science, social services, medicine and business. In order to provide an environment where this supercomputing can thrive, network security is imperative. Computer clusters are now being connected by TAPs and Packet Brokers to provide robust protection without impacting network reliability or availability.