Edge Computing Self-Fulfilling Prophecies

With the caveat that it is unclear how many respondents participated, a survey published by BPI Network suggests nearly-universal deployment of mobile edge clouds for 5G networks, though it is unclear what specific roles are envisioned by mobile operators. 

“Virtually all respondents (99 percent) say mobile edge clouds will be either very important (38 percent) or important (61 percent) to realizing the promise of 5G,” the study says.  “In fact, some 65 percent of survey respondents say their companies plan to deploy mobile edge clouds on their networks within 18 months.”

The actual question asked appears to be “How do you view mobile edge clouds for more localized data hosting and processing to realize the promise of 5G?” The specific role envisioned by the mobile operator was not the point of the question, but only the role of edge computing as a value for 5G networks. 

source: BPI Network

This is a bit of a self-fulfilling survey prophecy. Since every overview of 5G emphasizes the three-fold value of extremely-high bandwidth; massive machine communications and critical machine communication (low latency use cases requiring edge computing), it is unsurprising in the extreme that edge computing, ultra-low latency or mobile edge computing show up very high and very frequently in stated values of 5G. 

source: NGP Capital

The bigger questions revolve around roles in the cloud computing and edge computing ecosystem, and what roles (and revenue) mobile operators can assume, beyond supplying connections. This remains an open question, but there already are signs the main role for a mobile operator is as a supplier of mobile or wireless connections, not the  edge computing function itself. 

"New Normal" for Cloud Computing Will Simply Look as "Normal"

Despite frequent predictions that “nothing will be the same” after the Covid-19 pandemic ends, it likely will be hard to quantify the impact over a longer period of time, say five to 10 years. That has been the case when looking at changes after the internet bubble burst of 2001, the Great Recession of 2008 or events such as 9/11. 

This forecast of hyperscale data center revenues from 2018 to 2024 might provide an example. One cannot actually see the impact of the pandemic. In some cases, it might take two to four years for prior trends to reassert themselves. In some cases it will not even take that long to hide the Covid-19 impact on time series data sets. 

source: Omdia

The point is that predictions that “everything has changed” or that a “new normal” will emerge post-Covid are likely to disappoint those who want those changes to happen. Underlying trends tend to reassert themselves once a disruptive event has happened, and passed.

Who Wins Most from Edge Computing?

Many firms hope and believe edge computing “will become a key strategic growth platform for the communications technology industry.” That is undoubtedly true, if we include the value of infrastructure products. “The business impact will be in the trillions of dollars,” estimates Futurion. 

source: Futurion

Since the 5G standard itself is built on a virtualized core network, 5G will lead to more virtualization of 5G networks. 

“The emergence of the edge could transform the public cloud by shifting the center of gravity away from public cloud, providing opportunities for cloud providers, communications service providers (CSPs), and enterprises alike,” says Futurion. But connectivity service providers face  challenges. 

The new ecosystem is complex, and connectivity providers will require skill to create new roles beyond access services. 

Technology areas particularly suited to benefit from the new edge cloud include content delivery networks (CDNs), distributed routing, edge networking, micro data center technology, smart Network Interface Cards (NICs), network integration and automation software; edge-compute gear; and edge-data management and operations.

SKT PIns 5G Hopes Substantially on Edge Computing

Mobile edge computing, also known as infrastructure edge computing, is provided by a mobile or other connectivity provider and is viewed as key to SK Telecom’s 5G growth plans, according to a study produced by STL Partners and sponsored by Intel and Dell Technologies.

That sponsorship suggests the clear financial advantage MEC has for computing infrastructure suppliers, but SK Telecom believes MEC also supports its 5G business in several ways. 

SKT believes it can move up the value chain by becoming a supplier of an edge computing platform, though perhaps not the actual computing function itself. 

source: SKT

MEC also reduces  5G backhaul traffic and core network load, providing lower costs to support video content delivery networks and application caching.

MEC is seen as optimizing end user and customer quality of experience by reducing latency, as well as enabling low-latency applications and devices supplying gaming and augmented reality or virtual reality experiences. 

SKT also believes MEC allows a more-strategic relationship with  enterprise customers, using edge compute facilities in its 5G network as well as edge computing on the customer premises.

IBM Goes to the Edge

IBM is stepping up its support for edge computing, announcing its new IBM Edge Ecosystem, including independent software vendors and global system integrators to help enterprises capture the opportunities of edge computing.

 

Equinix, Vodafone Business and Samsung already are working with IBM to support edge computing. 

 

The IBM Telco Network Cloud Ecosystem  already includes Cisco, Dell Technologies, Juniper Networks, Intel, NVIDIA, Samsung, Packet, an Equinix Company, Hazelcast, Sysdig, Turbonomic, Portworx, Humio, Indra Minsait, Eurotech, Arrow Electronics, ADLINK, Acromove, Geniatech, SmartCone, CloudHedge, Altiostar, Metaswitch, F5 Networks and ADVA as members. 

IBM also has built a number of new edge computing initiatives based on Red Hat OpenShift, the enterprise Kubernetes platform. The new solutions include:

• IBM Edge Application Manager – an autonomous workload solution supporting up to 10,000 edge nodes simultaneously by a single administrator. It is based on an open source platform.

• IBM Telco Network Cloud Manager – a new solution that allows service providers to manage workloads on both Red Hat OpenShift and Red Hat OpenStack Platform.

• A portfolio of edge-enabled applications and services, including IBM Visual Insights, IBM Production Optimization, IBM Connected Manufacturing, IBM Asset Optimization, IBM Maximo Worker Insights and IBM Visual Inspector. All offer features to give clients the flexibility to deploy AI and cognitive applications and services at scale. 

• New dedicated IBM Services teams for edge computing and telco network cloud that draw on IBM's deep expertise to help clients deliver 5G and edge-enabled solutions across all industries.

Industrial IoT Impact

According to Nokia...

IoT and Edge Computing Still Pose "Dumb Pipe" Challenge

It is not news, but a recent Omdia survey suggests connectivity providers still primarily are viewed as having a trusted role as suppliers of internet of things connectivity, rather than platforms, service management or applications. In a nutshell, that is the “dumb pipe” role, where the connectivity provider’s role is the transport of packets over a wide area network. 

What might surprise you is that about 28 percent of survey respondents cite other participants in the ecosystem as trusted suppliers of the IoT connectivity function. Those trusted providers include system integrators, software suppliers and IoT specialist firms. 

source: BearingPoint

That is not as surprising as it might seem, as 5G and 4G mobile platforms now include support for private 5G networks, built using unlicensed, shared or reserved licensed spectrum. If use of licensed or unlicensed spectrum no longer is a barrier to creating a private premises network, as was the case for Wi-Fi, then it makes sense that IoT networks also can be envisioned much as local area networks are: using unlicensed, shared or licensed LAN assets to create private local networks.

The big difference is that private IoT networks might use local computing support--on the premises or within 10 kilometers in some cases, up to 100 kim in other cases--instead of backhauling traffic over the wide area network to remote data centers. 

None of that means enterprises running private 5G will dispense entirely with WAN connectivity. That would be necessary for other reasons. But it is likely that local IoT computing will do much processing locally, with WAN connections required for reporting and long-term storage, at the very least. 

The larger point is that IoT, like any other important use case, might often have connectivity providers largely viewed as suppliers of WAN connectivity, rather than platforms or apps, which is where most of the revenue and profit are likely to be created.

Microsoft Edge Computing On Premises, On Mobile Network, Standalone

As edge computing features a few different deployment architectures, including computing on devices, on enterprise premises or at local remote locations, so Microsoft’s edge computing solutions span those choices of computing location. 

Microsoft’s Azure Edge Zones “with carriers,” where Azure services and infrastructure are deployed  directly on 5G networks in the carrier’s data center, says Yousef Khalidi, Microsoft VP, Azure Networking. Functionally, that means less than a 10-millisecond delay between the cell tower and the edge computing location. 

Microsoft also will deploy its own standalone Azure Edge Zones in select cities, arguably more of a regional edge approach. 

Azure Private Edge Zones are  private 5G/LTE networks combined with Azure Stack Edge on premises, and representing the “on the enterprise premises” form of edge computing. 

source: Microsoft

On device processing has almost no latency. An on-premises server might have latency of a millisecond or so.

Edge computing at nearby locations in a metro area might have latency as low as a millisecond, or as high as six milliseconds, depending on where the processing happens in a metro area. 

Far end processing at a hyperscale data center has latency between 50 milliseconds and 55 milliseconds.

source: Deloitte

How Big is the Multiaccess Edge Computing Market, and Who Benefits?

In an era where applications can be provided to customers and end users over any internet protocol network, and since the global telecom industry has adopted IP as the standard for its public and private networks, markets have become more complicated.

Virtually all content, connectivity and commerce markets now feature business models based on sales of infrastructure, on-demand services, connectivity, real estate or transactions spanning many roles in value chains. 

So it now is harder to tell precisely what is meant when anybody talks about the size of a market. Consider communications provider multi-access edge computing. 

That market is projected by Grandview Research to reach U.S. $15.4 billion by 2027, growing at a compound annual growth rate of 38.6 percent, albeit from a small base.

What is tricky are the actual segments. Grandview notes the role of OTT media streaming services and personalized content, which will “drive the market growth.” 

That might otherwise be called the content delivery network business, which already exists, with the same range of potential roles MEC envisions. 

It is not clear to some of us that this is a useful definition, or even a new market. Of course, MEC is said to support many other use cases, such as various internet of things applications, which Grandview notes. 

It is possible that most of that projected IoT revenue grows for suppliers of cloud infrastructure and “edge computing as a service.”

Cloud Computing Growth Remains Steady

The easy temptation when looking at business performance these days is to attribute results to the impact of the Covid-19 pandemic. Sometimes that is appropriate. Shutting down an economy will have dire and direct impact on businesses and employees. But not every trend is affected foundationally.

Cloud computing has been growing steadily for years, and is projected to keep growing for the foreseeable future. Over a decade’s time, we will likely be unable to tell, just looking at growth rates, when recessions or periods of economic growth happened. 

source: Synergy Research