What goes around comes around in ICT, and edge computing has been in vogue for some time now, though with actual business cases and value chain models for telcos remaining uncertain. A new global initiative called Seamster, centered around Deutsche Telekom’s existing edge spin-out MobiledgeX, has been formed to try to address that lack of clarity by focusing particularly on use cases for edge computing alongside 5G.
This is addressing a new demand that has arisen from several trends. Initially, the rise of the cloud centralized data processing in big server farms. Two key drivers of distributing some of those resources closer to the user, at the network edge, were the growth in low latency applications and the proliferation of data generated towards the periphery of networks rather than in big enterprise centers. This data in turn increasingly requires processing locally, either for pre-preparation or to feed analytics, even if it ultimately migrates back towards the center, with edge computing reducing load on bandwidth as well again as reducing delays.
Stricter privacy regulations spearheaded by the EU’s GDPR, whicht came into force in May 2018, have also encouraged edge computing in some contexts by imposing rules over where data can be stored and processed.
5G has now added further fuel to these trends because edge computing is integral to many of the associated use cases and essential to avoid overwhelming backhaul networks as more and more data is generated down at the RAN level. 5G deployments will accelerate mobile data generation, leaving backhaul infrastructure struggling to keep up, at least for a while, with edge computing offering relief.
More than that, the capabilities enabled by 5G will encourage new applications that will require edge computing in order to deliver on the promise within a number of industry verticals. It was 5G and the prospect of these new applications that spawned Seamster, which aims to define use cases and so spur adoption of 5G-enabled edge computing.
Seamster is made up of mobile edge computing specialists such as MobiledgeX, which seems to be in the thick of most such initiatives. Indeed, there seems to be a plethora of such initiatives, raising the question of how they all fit together and why another is necessary.
MobiledgeX itself has been involved in various proofs of concept and then only in February, a few weeks before Seamster’s launch, a group of leading operators backed by GSMA unveiled the Telco Edge Cloud as an interoperable platform for edge computing, aiming to make applications and services more readily available to enterprises primarily at this stage, although eventually also end users. This is a partnership between operators China Unicom, Deutsche Telekom with MobilegeX, EE, KDDI, Orange, Singtel, SK Telecom, Telefónica and TIM, which all pledged to develop the platform during 2020, although progress is likely to be impeded temporarily by the coronavirus crisis.
The plan was to deploy the platform first across Europe, before extending globally later in the year or 2021. The main point is to share some infrastructure as well as expertise in latency, computation and storage with relevant software vendors, enabling them to meet the needs of their enterprise clients.
There are also various other edge computing initiatives less specifically focused on mobile operators, such as the European Edge Computing Consortium (EECC) supporting enterprises – especially smaller ones – that want to adopt edge technologies. There is also the Open Edge Computing Initiative (OEC), again involving a number of operators and technology giants such as Microsoft and NTT, as well, once again, Deutsche Telekom with MobiledgeX. This is focused on technology challenges tackled through an edge computing test center with the help of academic partner Carnegie Mellon University.
Against this background it was not surprising that, during the virtual event for its launch, Seamster felt the need to justify its existence, especially as many of the partners were already involved in some of the other initiatives. What is somewhat different from some of the other efforts is the number of partners which are associated with mainstream enterprise IT – and in some cases, pioneers of virtualization, which helped clear the path for cloud computing and then the procession towards the edge. These include Dell EMC, VMware, Canadian network monitoring provider Accedian and supply chain specialist World Wide Technology (WWT).
It was Gabriele Di Piazza, VP for solutions and marketing at VMware – the first major company to extend virtualization down to smaller Intel-based servers outside the realm of big data centers – who addressed the question of why there was a need for yet another global initiative in a field that has been established for some time in the enterprise.
The main answer was 5G and associated new applications based on mobility. This gelled with the drive for greater automation around cloud computing, which itself is becoming partitioned between the center and the edge according to where the best place is to store and process data. There was a need identified by the Seamster partners to relate these developments to emerging 5G technologies and focus on how these affect individual use cases. Each sector varies over the optimum balance between edge and centralized computing and storage.
Industrial IoT (IIoT) is one sector where mobile edge computing is primed to make an early impact, partly because many manufacturers are on the verge of upgrading to a new level of automation as robotics aided by machine learning progresses beyond just preset repetitive tasks. At the same time, ML is underpinning advances in preventative and predictive maintenance that will rely on a combination of centralized ongoing analysis of machine data and localized processing of real time diagnostics to optimize operation and warn of impending issues on the manufacturing floor.
Autonomous driving is an incipient sector where edge computing will make an impact, even if there are disagreements over the extent of it. In this case, edge would be in the middle of a three-layer architecture between the cloud and in-vehicle computation.
Clearly all safety critical processing associated with the immediate driving tasks such as steering, breaking and acceleration have to take place inside the vehicle to meet that definition of autonomy, but processing of data for, say, real time navigation could best take place at the edge by roadside systems, reducing latency and backhaul load.
This is where the greater capacity and speed of 5G will come in, being far better placed than WiFi for this use case because it can more readily provide full coverage spanning major road networks. This case raises different architectural considerations from IIoT where edge computing would play a more primary role in the most critical processes.