The partnerships between leading cloud providers and telecom companies, including one announced earlier this week involving Google and Spain’s Telefónica, to deploy multi-access edge computing (MEC) infrastructure at the edge of mobile networks indicate an eagerness to exploit the new era of 5G, which for the industrial world promises to usher in the flexible installation and use of technologies such as augmented and virtual reality (AR/VR), machine vision, robotics, and real-time analytics.
The ultra-low latency required by smart factory applications is enabled by eliminating the time delay to/from cloud-servers based in distant data centers through the provision of edge processing at a more proximate location. It is instructive to take a closer look at the edge approaches and technologies from Amazon, Google and Microsoft for 5G networks
Amazon: AWS Wavelength
Among the slew of announcements last December at the week-long Las Vegas extravaganza that is AWS re:Invent was that of the launch of AWS Wavelength, which promises single-digit millisecond latencies for applications including machine learning inference, augmented and virtual reality as well as live video streaming and gaming. Telco partners for AWS Wavelength include Verizon, Vodafone, SK Telecom (Korea), and KDDI (Japan).
AWS developers deploy their applications to Wavelength Zones, which are infrastructure deployments that embed AWS compute and storage services within telco data centers (i.e. at the edge of the 5G networks), and which can access the suite of AWS resources such as AWS Identity and Access Management (IAM), Amazon Elastic Compute Cloud (EC2), etc.
What this means is that factory data such as machine vision images can be processed by servers in Wavelength Zones without needing to leave the telco’s network. Without Amazon’s MEC, applications would need to get onto the internet to access AWS resources, thus adding latencies of 100 ms or more and discouraging the deployment of ultra time-critical applications.
Google: Global Mobile Edge Cloud (GMEC)
In March, Google Cloud unveiled its Global Mobile Edge Cloud (GMEC) strategy, which aims to deliver a portfolio of 5G solutions developed jointly with telecoms companies on an open cloud platform and deployed on a global distributed edge. That announcement also came with news of a partnership with AT&T to help enterprises, including those in manufacturing and transportation, to leverage Google Cloud technologies such as AI/ML and edge computing using AT&T 5G network connectivity. According to AT&T, combining 5G with Google Cloud can unlock the cloud’s true potential and provide the opportunity to create new experiences for customers.
Since then, Google Cloud has inked further telco partnerships, with Telefónica in Spain and Orange in France. The collaboration with Telefonica, announced June 11, aims to accelerate Spain’s digital transformation and advance 5G mobile edge computing by combining Telefónica’s cellular network infrastructure with Google Cloud’s mobile edge computing platform to deliver 5G based services to Spanish businesses. And just this week (July 28) Orange and Google Cloud announced a wide-ranging strategic partnership that includes the development of edge computing services for Orange’s 5G networks, which are being rolled out across Europe.
Microsoft: Azure Edge Zones
This March also saw the preview launch of Azure Edge Zones from Microsoft, which it touts as a transformative advance to combine the power of Azure, 5G, carriers, and technology partners and enable new scenarios for developers, partners, and customers. As with the other cloud providers’ offerings, the aim is to deploy cloud services into edge computing infrastructure to address applications requiring low latency and high speed and bandwidth.
In Azure Edge Zone with Carrier, one of the three types of Azure Edge Zones, Azure services are connected directly to 5G networks in a telco’s data center. Microsoft estimates that the single network hop from edge-deployed Azure to the 5G network will allow for latency of application-to-device of just 10 ms. Aside from AT&T, Microsoft’s original collaborator for the initiative, telco partners include NTT Communications, Telstra, SK Telecom, and Vodafone.
As its name implies, Azure Private Edge Zone is an on-premise version of Azure Edge Zones that supports a customer’s private cellular (5G or LTE) network and also allows for security function implementations such as firewalls. Users get low latency access to computing and storage services deployed on the premises of e.g. a factory. Offshore operations and security isolated facilities are also among the use cases for Azure Private Edge Zones.
The third type of Azure Edge Zones, simply called “Azure Edge Zone”, is a small-scale extension of Azure deployed in Microsoft-owned data centers located in urban locations and usually far away from Azure Regions. While there is no integration with cellular networks, which eliminates use cases such as mobile robots and connected vehicles. Azure Edge Zones support VMs, containers and selected Azure services to enable low latency and high throughput applications close to customers facilities.