There is nary a commercial or advertisement that doesn’t remind the public of the most exceptional promise of 5G - its ability to provide lightning fast speeds that will transform the entire communication industry, providing enhanced connectivity for a number of consumer services such as video streaming, browsing and more.
However, as 5G increases the amount of bandwidth, there also comes a need to develop more complex infrastructures to support such seamless connectivity. This is where the mobile edge comes into play and is one area that is challenging operators - a significant barrier to effective and widespread 5G rollouts - not to mention the need for excessive tower count, limited space for hardware support as well as the accumulation of even more connected devices.
As the device count increases, the support system behind it needs to adapt without growing its own footprint. An example of this is the hand off. One of the major issues with 5G is both the distance and the amount of data that is required, thanks to all of the towers transmitting and moving bits of data all over the place - hand offs from tower to tower. As you stream a movie, or ride in the car and speak on your mobile device, streaming becomes more challenging since the distance between those towers is substantially reduced. The trick is to move significantly more data, and faster, while still reducing the amount of data movement for the given edge location.
And this is where a new technology called Computational Storage comes into play…
The edge of computing and the start of 5G
So, chicken or egg - which comes first? The edge or 5G?
The fact is, 5G is edge driven which in turn is helping to drive a data-driven economy that requires new storage and data management architectures. But edge computing still requires data movement, just over smaller distances versus traveling all the way to the cloud or to a data center hub.
Edge devices, from traffic sensors to smart cars, provide possibilities for harnessing this growing in-device computing capability in order to provide deep insights and predictive analysis in near-real time, is nothing short of revolutionary. The edge’s proximity to data at its source can deliver real business benefits: faster insights, improved response times and better bandwidth availability and its distributed computing framework brings enterprise applications closer to data sources such as IoT devices.
The boom of the Internet of Things (IoT), driven and enabled largely by the widespread adoption of 5G in 2020 and beyond, will undoubtedly lead to the growing necessity for more edge computing devices as this is far more efficient than connecting to massive cloud data centers that are centralized and have limited capacity to process data quickly due to latency issues.
As a matter of fact, Gartner estimates that by 2025, 75 percent of data will be processed outside the traditional data center or cloud.
But here is the rub: for edge platforms to deliver on the stringent demands of 5G, they need to adopt a new architecture with intelligence and the ability to have both computing and storage requirements. In addition, the simple fact that transmitting all this ‘raw’ user data amounts to a huge security risk and can be mitigated by doing local compute.
In fact, a Western Digital blog spells it out perfectly “…as the result of developments such as machine learning and artificial intelligence (AI), 5G-capable devices are going to be expected to have even more advanced computing capabilities than what’s found in today’s devices.”
How to process information quickly at the edge: computational storage
Edge computing provides the ability to process terabytes of information quickly at the source, but how quickly, will depend on the intelligent storage architecture that is powering the edge device.
Let’s take an autonomous car for example: an autonomous car produces approximately 5GB of data per second - which can mean terabytes of data per week and petabytes per year. Therefore, edge devices need a powerful mean to process all those bytes and that's where Computational Storage comes in - providing an intelligent option of managing and processing so much data on a daily basis.
Look at it this way: today we have a mix of hybrid and gas cars, with a myriad of data needs. For example, not all Tesla owners get the same update at the same time - why? Bandwidth and support infrastructure. And now imagine if every car were truly autonomous. The number of bits flying between car, edge, endpoint, cloud, compute, storage - is just mind boggling.
Computational Storage is an innovative new technology that decreases data movement, processing the data right where it is created - rather than having to move the information all the way to the host CPU or to a hub data center for processing. In some cases, like NGD Systems, Computational Storage comes in small form factors that provide an added compute power, that can pack an analytical uppercut punch for the limited-sized and power-enabled edge data centers that live at each of the 5G new cell tower platforms. This works well for edge devices that are small, cramped spaces.
Providing additional compute to the confined resources that exist at the edge is paramount to the successful growth of the 5G space. Instead of requiring even more hardware and power to the server, the advent of high capacity Computational Storage provides the needed offload to the system to allow for great deployments.
There are many applications of this new technology of Computational Storage that benefits 5G, the edge, and most importantly the users of the technology. From security camera footage being ‘monitored’ locally by the AI in the device to the gas pump authorizing your purchase by simply looking up your car’s license plate, as well as the ability to find a lost pet who is wearing a IoT tracking collar - all of these are solutions that can benefit from the tools being deployed and performing more compute at the local platform.
