- Edge computing can enable a wide range of technologies as part of a digital transformation program.
- Edge computing supports low-latency services and offers new ‘branch in a box’ capabilities.
AT&T’s 2019 Business Summit had a big focus on edge computing. AT&T offers a convincing vision with use cases for both its multi-access edge compute (MEC) and network edge compute (NEC) platforms. However, the questions from the enterprises present at the event suggested that many are uncertain about what edge computing means and what it may offer their businesses. This is unsurprising, as the majority of service providers are still developing their own models for deploying edge compute capabilities, let alone offering fully formed go-to-market models.
Given edge computing’s immaturity, why then should enterprises care? The answer is that, for those enterprises working on their digital transformation strategies, the potential benefits of and use cases for edge computing make it an important technology area. This is particularly true when it comes to supporting other emerging technologies such as AI.
What Does Edge Computing Mean and Do?
Edge computing is one of the telecoms world’s most self-descriptive bits of terminology. Edge computing is the deployment of compute resources closer to the customer site. But, importantly, it is co-located with virtualized deployments of the service provider’s core networks (e.g., the provider’s mobile and fixed networks). The benefit of doing this is that a larger range of functions (e.g., AI, security, WAN optimization) can be provided without the need for deploying equipment at the customer site. It also means latency-sensitive applications can be deployed closer to the customer site, thus reducing latency.
AT&T provides a good example of two different ways this can be done. Its MEC platform is designed to be deployed at the customer site. This is a more expensive option, but it offers ultra-low latency with roundtrip times as low as 5 ms. This will support use cases such as automated factories and a range of augmented reality scenarios.
The provider’s NEC platform utilizes cloud compute PoPs in metro areas (AT&T it will eventually deploy around 100 of these across the USA). These sites will offer latency of around 20 ms, but because they are multi-tenanted, they will be more accessible economically for branch sites and SMEs. Latency at 20 ms will support advanced services like AI solutions for verticals such as retail (sentiment analysis, facial recognition, and ‘magic’ fitting room mirrors). Furthermore, it will also support advanced ‘branch in a box’ solutions which remove the need for IT expertise on site.
So, edge computing may be new, but it is opening up a wide range of potential technologies that could offer both cost savings and competitive differentiators. As enterprises are considering what the future of their business will look like, they should be talking to service providers about their edge computing deployment plans.