High Performance Computing at the Edge and in the Cloud
By Brian D. Kelley, CTO, Ohio Turnpike and Infrastructure Commission
Cloud computing certainly has its niche, but just like real clouds, cloud computing is a good distance from the data center on the ground. IoT devices and sensors are close to the ground just like fog and require a different type of computing power dictating that data is kept much closer to the ground. This is where edge, or fog computing can be leveraged as data is created, processed, and acted upon real-time instantaneously literally at the edge outside of the data center. Gartner estimates that there will be 25.1 billion IoT endpoints by 2021. According to Gartner, “Much of the current attention on edge computing comes from the need for IoT systems to deliver disconnected or distributed capabilities to the IoT world.” Factors driving the momentum to move toward edge computing include latency and content. Users expect fast response times and content such as video is too voluminous to send to the cloud and store in the cloud. Edge computing addresses these challenges.
Edge computing is being leveraged now in IoT with advancements in connected vehicle technology connecting vehicles to vehicles (V2V), vehicles to infrastructure (V2I) such as stop lights, and vehicles to pedestrians (V2P). Connected vehicle technology will eventually lead to fully autonomous self-driving vehicles. Experts agree that connected and self-driving vehicle technology will significantly reduce the nearly 40,000 annual fatalities occurring in traffic crashes on U.S. roads. High performance computing (HPC) at the edge will make this possible.
HPC provides not only super-computing processing capabilities, but fast network performance, vast amounts of disk storage, and substantial amounts of memory for highly complex processing challenges
A real-life application of this technology has been deployed at the Ohio Turnpike where 241 miles of toll roads and fiber cross through Northern Ohio from Indiana to Pennsylvania. The Ohio Turnpike has lit up a 60 mile test section of the toll road with a dedicated short wave radio communications (DSRC) connected vehicle technology pilot project This pilot project leverages edge computing to send real-time operational messages from thirty-eight Ohio Turnpike vehicles outfitted with DSRC technology and conversely send messages to a human machine interface (HMI) installed in the vehicle. The HMI messages sent to the vehicles provide real-time data updates on weather alerts, construction zones, incidents, and curve speed to a display device installed in the vehicle. The Ohio Turnpike hopes to achieve safety and efficiency benefits by deployment of this technology at the edge utilizing HPC.
According to Technopedia, “High-performance computing (HPC) is the use of super computers and parallel processing techniques for solving complex computational problems. HPC technology focuses on developing parallel processing algorithms and systems by incorporating both administration and parallel computational techniques.” To put it simply, HPC aggregates computing power to incredibly high levels of performance far beyond what an average desktop workstation could deliver. HPC provides not only super-computing processing capabilities, but fast network performance, vast amounts of disk storage, and substantial amounts of memory for highly complex processing challenges. Such aggregated computing power with HPC is a game-changing technology for solving complex computational challenges in health care, finance, science and engineering, business analytics, cybersecurity, and many other disciplines. HPC is also driving the machine learning and deep learning technology evolution with artificial intelligence. Gartner predicts that by 2022, 40% of customer-facing employees and government workers will consult daily an AI virtual support agent for decision or process support. HPC will provide the computational power, memory, storage, and networking performance to make this happen.
We have examined HPC’s critical role with edge computing far from the cloud, but HPC also provides amazing new capabilities in the cloud as well! Running HPC in the cloud provides the capacity to launch or scale up high performance computing clusters as needed, where needed, and when needed. Such HPC clusters provide an extremely flexible and elastic architecture that can be instantaneously provisioned on a per project basis in the cloud where you only pay for what you use. HPC in the cloud is an extremely attractive offering which significantly accelerates both productivity and achievement of desired results at a reasonable cost.
In the digital age of 21st century HPC will provide the computational power and capabilities that we can harness to solve complex problems in the world to drastically improve our health, well-being, and quality of life at home, at work, and when we travel. A plethora of amazing possibilities powered by HPC wait just over the horizon both at the edge and in the cloud.