The perils of poorly designed edge networks
5G traffic will include a significant proportion of communication between devices and between devices and people. To seize those 5G opportunities, service providers must distribute their resources to the locations where they provide the greatest benefit.
I recently ran an experiment that demonstrated the perils of poorly designed IOT networks.
I was mixing a healthy sports drink in a blender my wife had just bought. The blender had a cord with an electrical plug, a single Power button, and a tricky locking mechanism. Since this was the first time I'd used it, I was unsure whether I'd locked the mechanism correctly. The blender would not turn on if the mechanism was not locked in place.
Without much forethought, I decided to test whether the locking mechanism was set correctly by pressing the Power button. As my finger reached for the Power button, my eyes noticed that the blender had no lid. And that it was full of water and powder. Unfortunately, I could not send a Stop signal to my finger fast enough, and the blender spewed powder and muck all over the kitchen counter, my shirt, and my face.
Had I been able to distribute processing resources closer to the action I would have been able to avoid both a mess in the kitchen and my wife’s unrestrained laughter.
According to the November 2018 edition of the Ericsson Mobility Report, by 2024, 25% of mobile data traffic will be transported across 5G networks. That 5G traffic will include a significant proportion of communication between devices and between devices and people. To seize those 5G opportunities, service providers must distribute their resources to the locations where they provide the greatest benefit. For instance, if an application such as an autonomous driving sensor required low latency, you would place resources close to the application. If a different application required low cost but had no low latency requirements, you would host its workloads in the datacenter, where costs are lower.
Applied to my enhanced mobile network involving the blender, that distribution of resources might look like this:
- Local compute and storage resources placed at my eyes so I could perform data analysis to determine which signals required high priority responses. This approach would be faster than if I sent those signals to the brain for data analysis.
- Those resources would send any high priority responses to the appropriate limbs without the need for approval from the brain.
- Resources at any of the points between my shoulder and my knuckle to process, as effectively as possible, high-priority signals such as <blender_full> + <no lid> = <abort!> that might come from my eyes or ears. Processing as effectively as possible could mean transmitting a high priority signal with low latency, processing lower priority signals with low cost, or something else.
That’s what Ericsson Edge NFVI was designed to do: to help you place your resources where they will do your applications the most good, and host your workloads on the resources that make the most sense.
The compact, cost-efficient design of Ericsson Edge NFVI also makes possible unified management of cloud native applications and virtual network functions (VNFs) running on the same platform. Using a market proven system-verified design and telco grade robustness, Ericsson Edge NFVI is ready to be rapidly deployed in demanding telecom environments. It is part of an end-to-end managed and orchestrated solution for distributed cloud that is remarkably easy to deploy and use.
Want to know more about harnessing the edge opportunity for 5G? Read our Edge Computing and 5G Report
For more information, download our Solution Brief on Ericsson Edge NFVI