It's Time to Separate The Edge From 5G

TL;DR/Exec Summary

• It’s become increasingly common to assume that 5G and Edge (or, “5G/Edge”) are reliant and dependent on each other.
• Edge and 5G are attempting to accomplish similar architectural goals, but focused on different software ecosystems, buying centers, use-cases and operating models.
• Edge solutions will benefit from 5G, but the solution space is much larger than 5G access networks.
• 5G will be a key enabler of Edge Solutions, but 5G has many use-cases that do not depend on edge computing.
• There is benefit to consider these as separate solution spaces with defined integration points, rather than a single solution space.

Earlier this summer, I was recognized as one of the "100 most influential people in the wide world of 5G" (#5G100) by Informa Tech for contributions that help "cut through the industry noise." It has long been my perspective that the industry is better-served with well thought-out perspective rather than marketing hyperbole, and that the credibility I've built in Telecom has been developed through candid dialogue on re-framing problem spaces, rather than only providing solutions.

For the last decade, edge computing has become an increasingly relevant and important conversation for the future of enterprise services and cloud architectures. As a foundational discussion:

1. the ability to localize applications to reduce service latency (for those applications where speed to decision-making and operations safety decisions have justifiable business cases);
2. alleviate variance/jitter; resilience/survivability of operations; cost-avoidance (due to localization vs transmission of content);
3. improve data collection and analysis;
4. and develop new experiences that rely on those characteristics

has been a forcing function to rethink the needs of the cloud developer community. To that end, the industry dialogue and solution development around edge computing - both the decentralization of traditional cloud from hyperscale, centralized data centers to more regionalized (local) facilities, and the movement of cloud computing on-premise - have been important for media, retail, manufacturing, healthcare, amongst other industries. 

While this solution space is important, there has also been a trend of appending “/Edge” to other developing industry trends to both try to broaden the conversation on edge computing and to find new opportunities and use-cases, especially in #telecom network modernization. It is this “/” trend - the inclusion and attachment of another industry trend to the edge, such as “IoT/Edge” or “5G/Edge” - that has complicated the solution space as a whole, amongst vendors looking to productize offerings, Telecommunications Service Providers defining their future-looking architectures and business priorities, and within the cloud service provider ecosystem as developers open discussions on edge computing.

Edge computing has some interesting intersections with other industry trends, but not dependency. As such, edge computing will continue to develop as an independent solution space in which Telecommunications Service Providers are likely to be one of the beneficiaries. For that reason, it may be worthwhile to separate “The Edge” from 5G, recognize that the two trends are symbiotic - highly-aligned in objectives but loosely-coupled in practice - and eliminate the confusion of trying to link the two. At the core of both of these initiatives is the need to decentralize functions, manage infrastructure in new ways, integrate with the public cloud, and provide a seamless experience for developer ecosystems.

This blog will look at the two trends and highlight why they are uniquely different yet often combined. 

The Rise of Edge Computing

Industry focus on the role of the “Edge” is not new. It is a paradigm that has existed in Internet architectures, public cloud, and networking under many different names.

Internet architectures: Decentralization of content and applications has been a consistent trend in the industry for the entire life of the Internet. Any time the “law of large numbers” applies - where the amount of anything (content, sites, devices, requests, etc.) is larger than the ability of centralized facilities to handle those requests, we have decentralized parts of the user experience. For instance, Content Delivery Networks have existed for many decades, and built a distributed DNS architecture for the global Internet. The industry decentralizes, because decentralization is the most logical means of scaling. 

Public cloud: In the early 2010s, as an increasing number of IT workloads started to migrate to public cloud infrastructure, the paradigm of “hybrid cloud” developed, allowing portability of workloads between on-premise, private cloud instances and public cloud environments. In essence, the on-premise portion of a hybrid cloud architecture was the “edge,” but the lack of a common control plane and bespoke integrations between enterprise IT vendors and public cloud providers made these hybrid cloud offerings unmanageable.

Networking: Interworking between the service provider edge and the customer premise edge has been a consistent source of revenue and innovation for the SP Networking industry. Whether it be the early days of circuit-switched networks, evolving to MPLS and introduction of both the provider edge (PE) And customer edge (CE), or the evolution of consumer connectivity and the role of customer premise equipment (CPE) and home gateways, the networking industry has long-thrived on finding ways to introduce innovation at the edges of networks, especially those facing their customers.

So why, then, has the edge become both a significant industry trend and a new battleground between cloud providers, infrastructure vendors, Telecommunications providers, and other facilities-based providers (ie, CoLo facilities)? For two core reasons:

1. Technology innovation: New technologies that allow more flexible hybrid cloud architectures - such as those provided by AWS, Azure, Google and VMware [alphabetically]; and flexible any-to-any networking - such as SD-WAN, have simplified the realization of the multi-cloud.
2. Business Model Innovation: Data-driven business models for just about every industry has brought with it both an explosion in the size of data sets and a drastic reduction in the windows of relevancy of that data. With any big data environment, it is typically more efficient to move compute and storage to the source of data, than to move data to where compute and storage reside. 

Enterprise use-cases looking to tap into the capabilities and developer ecosystems of the public cloud providers that require significant compute/storage resources, and rely heavily on networking and data, including AR/VR use-cases, predictive maintenance, logistics (movement), etc. are all well-suited, across many industries, for edge computing. Multi-access edge computing (MEC) is a clear example of such an industry trend targeting this intersection.

Certainly, evolutions in the access network play a major role in enabling the edge. SD-WAN, 5G access networks, which provide higher capacity, lower latency, and finer granularity (for tuning capabilities to specific use-cases), and private mobility networks based on CBRS, which provide better manageability and experience compared to WiFi networks, have emerged as foundational enablers of both the technology and business-model innovations driving the need for edge computing. It is for this reason that Edge and 5G are often building blocks requiring interworking within the same solution space, but the edge solution space is also built on non-5G access networks.

The Role of 5G

In parallel to the edge computing industry dynamics, defining 5G access networks have been the focus of standardization organizations since 2015 (in 3GPP), with recognition that 4G networks would be insufficient for a number of reasons:

1. Consumer demand for higher-quality interactive video and AR/AR gaming requires higher bandwidth and lower latency
2. Increasing number of use-cases at vehicular speed demands faster mobility
3. Increased number of “things” connecting due to the growth of IoT
4. Mission-critical traffic, such as first responders, required higher reliability

In the end, the definition of 5G has been built around the “law of large numbers,” leading to the emergence of a new architecture, led by Telecom network vendors, that is decentralized, by-nature. This has led to the emergence of the perception of new “edges” in 5G architectures, but should instead be thought of as the same logical edges as legacy 4G architectures, instantiated at an increased number of facilities, augmenting the number of localized connection (peering) points to the public cloud. As an example, the 4G Evolved Packet Core (EPC) architecture relies on tunneled packet termination at the Packet Gateway (PGW) much the same way that the 5G Core relies on tunneled packet termination at the User Plane Function (UPF). 

In addition, the Telecommunications trends for disaggregation of both hardware and software, and of control and user planes, has led to more cloud-like network architectures driven by open, API-driven interfaces, cloud-based control points, decentralized network functions, and DevOps-style operating models. In the end, however, 5G architectures are defined by the (growing) Telecom vendor ecosystem, focused on operations and network functions themselves, and entirely driven around improving the capabilities of the Telecom network. To summarize, the 5G architecture has spurred a need for modernization of the entire Telecom network. 

So, Why The Consistently-Combined Messaging?

The combinatorial messaging and the urge to integrate these two trends versus isolating them as unique market inflections, seems to stem from the need to generate demand for 5G networks. The Telco demand for edge is a financial demand not an architectural demand, driven by the need of the Telco industry to find high-margin value in the deployment of 5G connectivity at the same time that “edge” is actively searching for use-cases to justify it. 5G needs to be monetized, use-cases drive monetization, vertical industries are driven by use-cases, and the “value” proposition of 5G is driven by the real-time, low-latency, interactivity, capacity, large numbers of devices, etc. - all of these are inherently similar to the use-cases that drive “edge”.

Further overlapping to the paradigm is the potential trend for Telecom facilities, such as central offices and backhaul aggregation locations, to become hosting facilities for MEC use-cases in lieu of on-premise offerings, essentially building a next-generation distributed Colocation market. This new local market opportunity requires a highly flexible network underlay (diverse, meshed fiber paths) and network overlay (network slicing, SD-WAN, etc.), spurring a network investment cycle in local and metro networks using software-defined technologies.

Lastly, the integration of connectivity into on-premise edge offerings, in which Telcos provide offerings of Edge PaaS + Connectivity, provide a channel to application developers, and provide turnkey solutions to vertical industries, focused predominantly on a use-case driven approach has generated a tremendous amount of industry momentum at the same time that 5G networks are being built. It has become un-natural, in some ways, to NOT conflate these two industry trends.

The Time Is Now

It’s clear that the technology and market opportunities of edge computing and 5G intersect pretty extensively. In fact, there is a lot of commonality in the underlying and enabling technologies such that engineering an “edge stack” and engineering a “Telco/5G Network Modernization stack” may be similar endeavors, or at least share some common functional building blocks. However, the time is now to focus on the independence of these solution spaces, while simultaneously ensuring open interfaces and API-based integration points, allowing these loosely-coupled offerings to be composed into unique solutions. I look forward to continuing to help influence the industry in this direction such that both 5G and edge computing are successful and ensuring that Telcos can properly monetize these investments.