Data at The Edge: Cloud Computing, Edge Data Centers Drive Completely New Paradigm for More Distributed World

By Arti Loftus, Special Correspondent  |  June 23, 2020

Earlier this month, NetEvents inter@ctive held a virtual event entitled “What’s Hot in Networking & Analyst Views,” bringing together a senior panel of specialists for an enlightening discussion about the technology sector’s business outlook during the COVID-19 pandemic.

The event featured John Apostopoulos, VP & CTO for Enterprise Networking Business, Cisco (News - Alert), with a panel chaired by Jeremiah Caron, Global Head of Research & Analysis – Technology Group, GlobalData. On the panel was SVP Marketing, Networking NVIDIA, Kevin Deierling (News - Alert), VP of Marketing at Mellanox Technologies since March 2013, Kevin Deierling recently joined NVIDIA as SVP Marketing, Networking when they completed their acquisition of Mellanox (News - Alert), initially announced in March 2019.

As part of the event, many questions were posed and opinions were discussed, such as the big issues architects need to be thinking about as we move into a world where we have cloud edge computing.

“I think the big picture is that you're no longer really programming servers, the new unit of computing is a data center, and you actually are running a distributed application across the entire data center,” Deierling said. “And with the edge, it really fragments even farther so instead of a giant centralized data center that has 10,000 computers, you have 10,000 data centers at the edge, each of them with a smaller rack, but still performing all kinds of new functionality.”

Because everything is untrusted in the edge, Deierling explained that if “you’re going to be plugging in cameras and devices that you simply have no control over, there's no security guard the way there is in a centralized giant data center.  Making everything software-defined and hardware-accelerated and doing all of that together really requires a lot of horsepower and a lot of innovation like AI and machine learning because all of these edge applications are unsupervised. When we look at the way people use the applications today for IoT, it is mostly the phone, and they are curated by humans controlling what is happening. This is pretty slow; the response time that a human needs to respond is actually very slow compared to what intelligent machines can do with the right automation algorithms built in. What the world needs is autonomous AI machines with low-latency networking that both reliable and secure.”

Cloud-native networking has expanded on traditional, monolithic applications that cannot support millions of users and devices, Deierling explained. “This is by expediting enormous growth in scalability that is now obtainable with a container-based solution that dispenses with single computers which are unable to cope with large scale issues. New cloud-native networking will be defined as everything needs to be automated - coming and going dynamically. Replacing a monolithic app, ephemeral containers connect between overlaying networks such as Virtual Extensible LAN (VXLAN). This permits for dynamic movement that can be used to build and scale up the data center utilised to run your application.”

As an example, Deierling said, “Maybe you have something that needs a ton of data so you need more storage connectivity and doing that dynamically with containers that are coming and going all the time becomes simply unmanageable, even with that very smart IT guy sitting in the corner. The future is in the rise of autonomous systems delivered on secure, high-speed networks.”

Deierling went on to explain his vision for the future of AI saying, “Nvidia is pushing AI and closing the loop between telemetry and action. To do this we talk about something that we call ‘time to innocence.’ Figuring out the time to innocence is intrinsic in uncovering the underlying telemetry. Using Mellanox’s newly introduced ‘what just happened’ (WJH) you can feed a ton of data through application and switch basics, while being very specific about the actual cause of any underlying packet drops or latency issues, etc. You then can close the loop with AI.”

“There are platforms out there that we have like the EGX, which is an edge platform that has a GPU capable of running 5G networks, codec stack, and closed loop. I think the world is realizing that AI is not a threat to the technologist, but rather it is a powerful tool. This is a matter of education. But smart CIOs and smart IT guys are not afraid that they're gonna lose their job, they just need to do some hard work to understand how to use these advanced new tools that are available. When I work with IT administrators, it is clear they are focused on two new technologies, one of which is the cloud. I think ultimately, they're starting to understand that cloud is fundamentally a better way to develop and maintain software.”

California-based Nvidia Corporation is known for its designs in graphics processing units (GPUs) for both the gaming and professional markets, and more recently for its system on chip units (SoCs) for the mobile computing and automotive market.

While expanding its presence in the gaming industry, NVIDIA (News - Alert) has been diversifying its business portfolio since 2014 to include professional visualization, data centers, and now artificial intelligence (AI).

“This is largely due to the revelations in GPU deep learning which is said to be the next era of computing,” Deierling explained to the virtual audience assembled for the annual Spring event. “With the GPU acting as the brain of computers, robots, and self-driving cars that can perceive and understand the world in its environment. With Mellanox, the new NVIDIA has end-to-end technologies from AI computing to networking, full-stack offerings from processors to software, and significant scale to advance next-generation data centers.”

Deierling has contributed to multiple technology standards and holds over 25 patents in areas including wireless communications, error correction, security, video compression, and DNA sequencing while serving as VP of Technology at California-based biotech company, Genia Technologies, Chief Architect at Silver Spring Networks (a subsidiary of Itron), and head of marketing and business development at Spans Logic, a chip company that was acquired by Cisco.


Arti Loftus is an experienced Information Technology specialist with a demonstrated history of working in the research, writing, and editing industry with many published articles under her belt.

Edited by Maurice Nagle