Welcome to the thirteenth in the controversial question series. A little delayed, but here we are.
In this issue, we were interested to hear the thoughts of senior industry executives as to whether they thought telcos would make a play to take on this role.
Here's the question we posed. "With the rise in both the use and importance of Edge Computing, will Telcos find a way to become important players in this resurrected communications-centric future?"
Glenn Agiran, associate director, telecom sales at Vertiv A/NZ kicks things off, saying that "As Australia prepares for a digital-led future, we've seen an inordinate adoption of technology in the last 18 months, partially due to the pandemic's impact. From the hybrid workplace pick-up to telehealth and the entertainment streaming boom, human-to-human interactions have decreased and digital has stepped up for the heavy lifting. Telcos by and large have shown that not only were they able to deliver the digital outcomes placed on them, but also demonstrated how critical telecom infrastructure is to the communications-centric future - because 'the frozen Zoom face' simply won't do."
Jonathan Jackson, director of engineering, APJ at BlackBerry continues a similar theme. "The enterprise is also moving to the edge. The rise in remote working and the swift march to cloud computing means corporate resources are brought outside the traditional network perimeter, and closer to the user and device. Everything, including data, apps, workflows, analytics, and decision-making, is being decentralised and moving to the edge."
"While cloud-centric computing is the direction the world is heading with providers such as AWS and Azure, edge computing is still a vastly relevant topic as it's all about ensuring efficiency in the way that computing resources are used," Osh Ranaweera, connect business manager at Somerville, suggests. He continues, "Edge computing ensures that you place workloads locally for the best use of resources, enhanced performance and lower costs."
Laying it out plainly, Loren Long co-founder and chief development officer - DartPoints, tells us that "The Edge is about participating in an ecosystem without seeking control or leverage. It's about developing and sharing resources that benefit all the players for the ultimate goal of improving how end-users experience content and application, whether human or machine. Carriers play a crucial role in this ecosystem, but it will depend upon their ability to collaborate with all of the other players. The days of owning and controlling how their subscribers connect and consume data won't work in a highly disaggregated world."
According to Naresh Singh, research director, Gartner, "Gartner forecasts that the edge hardware infrastructure market will be worth US$17 billion globally by 2025. The corresponding software and services opportunities will be even higher. However we expect that by 2025, less than one-third of telcos will offer their own edge cloud compute services for customer applications. The others will tactically resell hyperscale partner offerings and miss out on revenue.
"Telcos also need to carve a strong leadership position for themselves in this highly attractive market. If they can design a differentiated edge compute offering based around the low latency and highly contextualised offerings that they can provide due to their proximity to edge locations, and those that make sense for the kind of customers they already work with, they will do well. Such edge capabilities could include IoT platforms, smart personalisation, advanced virtual assistants, Edge AI, tokenisation, smart contracts, distributed cloud, advanced computer vision… The choice of opportunity should be on the basis of what makes sense for the kind of customers that CSPs already work with and those that can actually bring tangible business value to the table."
So, where is the Edge?
Of course, our introduction was quite vague in how it defined 'the edge,' - somewhat deliberately as we wanted the executives to offer their own thoughts. No-one really took the bait!
Stephen Gillies, technology evangelist at Fastly suggested that "The 'edge' has different meanings for different people. For a telco the 'edge' could be the mobile tower, as it identifies the hand off point between their networks and a customer device. For content providers and the vendors who support their delivery such as Fastly, the 'edge' is the Content Delivery Network - the hand off point between the streaming media application (often a browser or media player) and the consumer's device. So, edge computing can be confusing depending on who you talk to! Is it Web Application Edge or Networking Edge?"
Vijay Kolli, regional vice president, enterprise security Asia-Pacific at Akamai Technologies essentially agreed with Gillies, offering, "The question telcos have to ask themselves is where the edge compute platform should be deployed - at the towers, the middle of the network or at the core. Where it should be deployed should be dependent on the popularity of the applications, and today we are seeing it to be at the core of the network."
This suggests that the tower operators have a role here
We almost felt as though both Kolli and Gillies were suggesting that small server rooms might be swinging half-way up the 4G and 5G towers!
With something of a vested interest, Eric Watko, vice president, innovation product line management at American Tower broadly agreed. "The convergence of cell towers and edge computing is a natural progression of the next-generation network evolution. Utilising existing communications real estate, where connectivity already occurs, tower companies can develop edge data centres on a large scale very quickly. Today, we see small- and medium-sized enterprises using our edge facilities to optimise network elasticity—the ability to quickly scale computer processing, memory, and storage resources to meet changing demands.
"We are also seeing hybrid approaches, where enterprises are using both edge and cloud-based services to gain the advantages of both models. The next phase of edge utilisation will be by Mobile Network Operators (MNOs) when 5G and IoT applications are more widely adopted, as it will rely on a more distributed architecture to lower latency. Open Radio Access Network (O-RAN) will be the driver of this future wireless ecosystem, leveraging shared Distributed Units (DUs) and Centralised Units (CUs), which our collocation Edge Data Centres are a great fit. Telecom companies are well-positioned to play a pivotal role in this shift."
Of course, with so many disparate players in the chain between our home and the iTWire web site, any successful play will require cooperation from many unrelated players.
Singh suggests that "The most favoured approach to edge computing by telcos is to tie up with strategic cloud providers like AWS or Microsoft. A number of agreements have already been signed around the world including Australia. But if this is the only strategy, they are making a big mistake. It may turn out to be a hole that's very difficult to get out of, and only accentuate the perception of telcos as just network providers and not strategic digital enablers."
"Let's not also forget that with successful deployment, there will be challenges around scale and security," Kolli reminds us. "Edge computing requires better delivery networks and infrastructure, as well as interoperability between mobile networks and enterprise applications. Securing information both vertically and horizontally will also have to be addressed. For telcos, this is an opportunity to partner with other IoT providers to secure endpoint access across a variety of use cases."
The influence of 5G (and 6G?)
"By marrying the benefits of edge computing with the speed (and growing adoption) of 5G, telcos have an opportunity here to value-add," says Kolli. "This is especially the case in industries that rely on real-time intelligence such as fleet management and the transport sector, or industries that depend heavily on the use of private networks, such as manufacturing or, mining."
"With an ever-expanding open data world, connectivity is critical to pushing necessary real-time data feedback loops to understand customer experience. Telcos will, therefore, play a vital role in ensuring the seamless flow of communication and information exchange. When traveling from a rural location into the city, customers often complain about experiencing a massive difference between the two regions. This is a massive opportunity for telcos to support AgTech and Mining services with edge computing to those areas with unstable connectivity - this makes all the difference when companies are trying to drive customer retention."
Jackson takes a similar view, observing that, "One of the most exciting arenas for the application of 5G-powered edge computing is connected and intelligent vehicles. As 5G networks mature, we look forward to seeing the enhanced operational, security, safety and personalisation outcomes possible, not only in the automotive sector but in all edge devices."
Turning from the physical to the more abstract, Jackson continues… "Meanwhile, we are recognising the collaborative and scalable benefits of elastic workflows, which are enabled by the cloud. These capabilities will continue to be augmented by 5G on the edge. Moving workloads to the edge increases the bandwidth required, and telcos will be instrumental in creating the 5G fabric to support the hyper-scalability (and horsepower) required. The last 18+ months has shown us that we can deliver digital transformation and 5G will accelerate this further for a scalable, cloud-connected software platform that facilitates decisions on the edge."
Of course, there's always a negative view! "Telcos are betting on their 5G networks, which they think will hand them the edge computing market on a platter," says Singh. "Hence, they are procrastinating over taking a more assertive role in this market for 5G services to be first operational."
Which leads us to ponder…
Counterpoint - are telcos the right people for the job, anymore?
Taking a somewhat pessimistic view, Anton Kapela, co-founder & CEO for EdgeMicro suggests that "Telcos didn't retain the raw brilliance that many evolved to support - picking on ATT in 2021 versus the 1960s is like shooting fish in a barrel. As a case in point, ATT used to support primary systems research projects that led to the development of Multics, and all sorts of UNIX variants. With this work came oodles of then innovative, working code, and a host of counter-intuitive discoveries.
"These developments lead to a more efficient way to use communications capacity than circuit-switching systems the industry had grown to accept up to that time. Maybe the "innovation streak" started with Bell Lab's 1947 discoveries in semiconductors that led to the production of cost-effective and useful transistors. Somewhere along the line, ATT and its brethren seem to have lost their way: we still have something of a Bell Labs doing great work, but we haven't had any "hits" like these in a while. Ones that do stand out, and which sadly haven't seen their due "time in the sun" yet include Gerard J. Foschini's amazing insights into spatial multiplexing for radio systems."
Pointing to potential existential issues within the telcos, perhaps there's far too much of 'we've always done it that way.' Agrian suggests, "With all the transformation comes opportunity. We're at the forefront of 5G and with latency sensitivity at play, the data generated needs a processing home at the edge. But the rise in edge computing won't be the only critical success factor - telcos will need to consider challenging their identity crisis. The cost vs value creation issue has only been exacerbated by the latest digital shift, and in the decade ahead, we anticipate success will come to those who can transform in the most cost effective and synergistic manner possible."
Of course, no discussion of 'The Edge' is complete without IoT rearing it's ugly head.
Kolli paints it like this; "With the world becoming increasingly mobile, telcos have first access to end-users and their IoT devices. This means that they can create opportunities for future application innovations. Historically, telcos charged OTT players for bandwidth rather than looking for ways to provide value. We're now seeing a shift in the industry as telcos are realising the benefits of being a part of an ecosystem. The faster speed of 5G and computing capabilities at endpoints mean that telcos can capitalize on this and be a part of an ecosystem that provide value-added real time services and innovations to businesses such as identity, location and low latency capabilities, enabling the OTT applications of the future."
Similarly, Michael Ewald, director of technology, Contino offers, "The industry seems to think that 5G is enough to make the fully integrated IoT and Edge Computing infrastructure work. But it is not good enough - we need to think of 6G and telcos as an integral part of this ecosystem.
Remembering, of course that the 'S' in IoT stands for 'Security.'
So, where are the issues?
Thus far, we've taken a relatively high-level view of the issue, but Gillies wants to remind us that there are some specific technical issues to contemplate as well. "Shared spectrum networking such as WIFI and traditionally high network latency last mile technologies like satellite can have a negative impact on edge computing performance; the incapacious component moves from how quickly a request can be assessed by a server running in a public cloud somewhere, to how quickly the network can provide the response from a CDN edge. Fastly's C@E instance startup time is microseconds, yet the latency introduced by some last mile networks is in the tens or hundreds of milliseconds.
"Telcos are looking at finding appropriate last mile solutions which deliver the lowest possible latency, and minimising contention ratios around their networking nodes. But the end to end path of a web application or edge compute solution needs to be considered carefully to remove any potential bottlenecks."
Perhaps it's a stuttering evolution
Kapela wants to take the entire telco industry to task for losing their mojo.
"While we do have some inclusion of MIMO in modern LTE and 5G NR, it's not the 40-odd layers he and his colleagues had working in the 1990s. We have Thomas L. Marzetta, who while at Bell Labs, essentially did the heavy-lifting of discovering how we might use wireless evenly, yet more efficiently, with his invention of 'Massive MIMO.' This sort of primary physical-layer communications research is hard - not just anyone likes to think in vector calculus. It's not the most accessible or sexy stuff to work on, compared to dot-com captivors and names like TikTok or Facebook. That's easy to get behind, but these are all simply applications that ride atop the stuff of the real, actual, physics of it all: the network. Without it, we'd have nothing like we have today in wireless, to be sure, but there's a problem: we still don't have even a sliver of what we should have received from these discoveries.
"Why? Our wireless telcos (I'm looking at you, mobile network operators) don't know what they should be asking for. They trust their vendors (Ericsson, Nokia, et. al.) to take things "from the lab to the tower" - only the vendors really don't excel at picking winners, they excel at riding the economics of big-capital holders at said wireless networks to drive the equipment deployment and depreciation cycle, where we enjoy incremental improvements, sometimes anyway (ie. 2G to 3G, to LTE, and NR) - but we haven't gotten more efficient at the physical layers.
"In 2001, EVDO rev A delivered a few megabits of usable capacity to one mobile station at a time. Today, 5GNR can deliver more in total, but, we're not seeing the gains that we do see in Massive-MIMO systems. We should be seeing orders of magnitude more spectrum efficiency, with all that we've figured out about how to best use radio resources. Instead, we're seeing GHz wide channels and MM-wave distracting nonsense bearing the "5G" banner, which isn't going to radically change the network landscape for anyone much, really. Maybe a small number of people who are a few hundred feet away from a MM wave base station.
"What we should have seen instead: Massive MIMO deployments on sub-6GHz bands revolutionising the density and aggregate capacity of wireless systems. We should have seen the wireless network turn the economics of bit-hauling upside down. We shouldn't have to be connecting every last house up to GPON or DOCSIS on coax, in 2021. We should have had a wireless, efficient, fungible network future. We don't have that, but we could have it, maybe: telcos need to return to their roots and realise their initial priorities were good, market distractions be damned.
"Embedding "more cloud stuff" (looking at you, ETSI MEC) in the "Telco edge" is not the sort of stuff that revolutionises or catalyses the global-scale internet cares about. Fundamentally, customers of telcos want to connect to the network with the greatest value: the internet. All parties involved need to accept this viewpoint, align with it, and run, not walk, to redoubling efforts in fundamental research and development of advancements towards deploying the stuff we've already developed and which we know can work. That's how we make our communications-centric future happen. Not any other way."
Although with language a little milder, Singh agrees. "Telcos have been known to miss opportunities - cloud, unified communications and collaboration, for instance. If the same happens for edge computing, they will likely be pushed into gathering the spoils at the bottom of the value chain, stuck with an increasingly commoditised network and staffing play."
Bruce Lehrman, founder and CEO of Involta launches our wrap-up by observing that "Telecommunications providers will be just one link in the chain in a communications-centric future. Edge Computing is now a fully integrated and necessary part of any IT discussion. It's a fundamental part of meeting new demands for speed, latency, geographic reach and customer satisfaction. Strategic geographic locations for data centres will allow service providers to collect data at the edge. Computations can then be performed on that data and network assets can allow the distribution of that data in a meaningful way to businesses. Organisations across manufacturing, healthcare, and beyond are looking to leverage the most innovative opportunities across IoT, 5G, AI, and more. Those tools and applications thrive at the edge. We know that edge is the next growth engine. That makes the edge the place to be to drive the most value and ensure a competitive advantage, and it hinges on location, location, location."
Most of our correspondents eventually found themselves taking a positive view of the increased telco involvement at the edge.
Jackson: "Telcos will almost certainly become key players in the edge computing ecosystem. 5G networks will underpin the intelligent, connected future we are currently building - from IoT devices and connected cars, to smart cities and infrastructure. Those that bring this framework to market first will differentiate by providing new value in the marketplace."
"Agiran: "As they look to what should be left behind - from ageing fixed line infrastructure and legacy IT systems, to the challenge of cost driven consumers and OTT player cherry picking - telcos will also be challenged to look at the future they're heading towards. Whether it's fully integrated vertical strategies targeting horizontal market capture, or the more conventional network-performance strategies, tailoring to each based on the telco's market position and company DNA ought to be a significant consideration as they look to ensure player importance."
Singh: "Grabbing the full opportunity that edge computing offers is not possible without a role that goes beyond providing just the network. A more strategic approach does conflict with the business objectives of the cloud providers they partner with. Hence, they need to exercise tact and balance. The network, how much ever a commodity it has become, is still an advantage that a CSP has — as without that, none of the computing edge or cloud or the digital would be possible."
Ranaweera: "Edge computing requires connectivity, more specifically, reliable connectivity to the Internet or centralised systems located in the public cloud. It's important to pick a telco with the right set of tools and connectivity offerings that provide a high level of resilience, availability and an effective support model for higher SLAs. Telcos will need to look at connectivity technologies that incorporate diverse media, support higher capacity and utilise diversified physical infrastructures layout to ensure higher uptimes. Technologies such as 5G allow greater levels of diversity and capacity which are the key ingredients for a successful edge computing architecture.
Finally, tying up with a nice little bow, Ranaweera offers, "Provided telcos apply the correct technology stack and the approach at the right time at the right cost, they are well positioned in the market to benefit from the current trend of edge computing and become a key player in that market segment."
What's that song lyric? "Don't push me cause I'm close to the edge…"