Supercomputers, once used only for high-level engineering and research like genome sequencing, will soon be readily available to businesses via the cloud. Two of the biggest heavyweights in tech, IBM and Microsoft, provide their enterprise customers with access to supercomputers today, which means supercomputing will be a reality for every business before you know it.  

Supercomputers are machines with high performance computing (HPC) or big computing features such as burst processing, graphical processing, and cognitive applications. Burst processing is rapid parallel processing of massive data sets. Cognitive applications are the processes we associate with artificial intelligence (AI).

Computer processing speeds are measured in floating point operations per second (FLOPS). A typical server processes at gigaFLOPS speeds. A supercomputer processes at petaFLOPS rates, which is a million times faster. More than the raw speed it is the parallel processing that sets supercomputers apart.

In a recent article on TechRepublic, Evan Koblentz argued that supercomputers will soon be common office equipment.

Supercomputers, according to Koblentz, are “to ordinary servers what race cars are to street vehicles.”

Advancements in technology introduced in specialty areas, such as racing cars or splicing genes, are incorporated into every day use after they have been tested and vetted by high-end users.

This echoes the trajectory of bandwidth use. Initially only universities and the military used the internet. Businesses soon followed, and home internet came next. Organizations with large budgets are always the first to deploy higher speeds, but eventually that technology is available to anyone. Today, fibre to the premise (FTTP) internet, capable of essentially limitless bandwidth, is readily available.

Symmetrical networks support more bandwidth because the downloads and uploads are able to send and receive data at the same rates. In asymmetrical networks, the download speed is much greater than the upload speed which, in the business world, can create inefficiencies with processing or working with larger data sets. These asymmetrical networks are typically supported by traditional copper lines, which most telecommunication companies use, an archaic infrastructure that will one day be a thing of the past in business applications.

According to Koblentz, the cost of the equipment is not the only challenge to supercomputer deployment. Another issue is that supercomputers can take months to set up. This may explain why IBM and Microsoft offer supercomputing only via the cloud.

Microsoft is working with Cray, the producer of one of the world’s fastest computers, to offer cloud services powered by supercomputers via Azure, its enterprise cloud service. Azure customers can have access to a dedicated Cray XC or CS to run HPC and AI processes.

Cray computers will be set up in select Azure data centers. Jason Zander, corporate vice president at Microsoft Azure, believes this will yield unprecedented innovation. In his words: “By working with Cray to provide dedicated supercomputers in Azure, we are offering customers uncompromising performance and scalability that enables a host of new previously unimaginable scenarios in the public cloud. More importantly, we’re moving customers into a new era by empowering them to use HPC and AI to drive breakthrough advances.”

Microsoft’s enterprise customers will be able to run advanced analytics and modelling while connected to Azure in a fraction of the time.

IBM’s Watson is a similar service. Watson is best known as the machine that won Jeopardy in 2011, and since then it has become synonymous with AI. According to Trips Reddy, senior content manager at IBM, “the adoption and application of artificial Intelligence (AI) continues to accelerate at an exponential rate in modern businesses” and will soon power many aspects of business operations, from machine learning in manufacturing to human resource management.

In her article, Beyond the hype: The reality of what AI means for business, Reddy outlined real-world business uses for AI. These included teaching Watson how to handle customer service queries for financial institutions, voice-controlled assistance in cars, and even tax preparation.

All this increased processing speed is bound to impact the bandwidth requirements across business applications. To enjoy the benefits of rapid processing, businesses will need to move massive amounts of data into and out of the cloud. Fortunately, the fibre optic network infrastructure that can move massive amounts of data is accessible to businesses of any size.

Want to learn more about future-proofing your business with fibre?

For more information on our secure fibre connectivity solutions, contact your Axia salesperson at sales@axia.com or 1-866-773-3348.