For the past fifteen years, hyperscale cloud providers have revolutionized both the landscape technology and business models of Information Technology (IT) and Operational Technology (OT). From CapEx to OpEx conversion to software defined everything, the cloud has been the most impactful catalyst of the Digital Revolution, the Cambrian Explosion of Data and the global Internet of Things (IoT). However, there are some key areas where the cloud has historically struggled to drive the same outcomes: the edge – the parts of the world where the internet does not reach.
You Keep on Using that Word
Before we get into the world of cloud at the edge, let’s discuss the basic terms. First and foremost, the cloud is NOT someone else’s data center (though it has certainly been misused that way by some in the past). The cloud is most clearly defined by the National Institute of Standards and Technology (NIST) in SP 800-145 to be made up of on-demand self-service, broad network access, resource pooling, rapid elasticity and measured service delivered via software- as-a-service (SaaS), platform-as-a-service (PaaS) and infrastructure-as-a-service (IaaS). Second, what is “the edge?” To keep it simple, it can be defined as any environment at the boundary or beyond where network access exists today. Coincidentally, broad network access is a prerequisite for the cloud to exist, so naturally, cloud and edge have not always mixed well.
Pushed to the Limit
What are the limiting factors of cloud at the edge today and how should they be addressed? One popular approach to work around limited Wide Area Network (WAN) access has been to leverage the portability and composability of containers and container orchestration, most notably Kubernetes. This certainly answers the mail in terms of resource pooling, rapid elasticity and measured service. It also provides options for on-demand self-service and broad Local Area Network (LAN) connectivity delivered via PaaS, with all the trademark characteristics of the cloud. However, edge-deployed Kubernetes is still inherently limited by the hardware and infrastructure on which it is deployed, presenting unsustainable / infeasible logistics and supply chain challenges.
Another avenue that is perhaps more promising is the elimination or clouding of the edge altogether through the emergence of ubiquitous 5G. Telecommunications providers are in an all-out race to provide not only higher bandwidth to every corner of the globe, but also open architectures of traditionally proprietary Radio Access Networks (RAN) – and many are leveraging the cloud to do so.
Let’s Make a Deal
This presents some unique business opportunities between the hyperscale cloud providers and global telco. Similar to how Netflix improved global customer experience through Open Connect, cloud providers like AWS and Azure are partnering with major telco providers to host solutions like AWS Wavelength and Azure Stack Edge for Carriers as a service to the telecommunications providers. This presents an immediate cost advantage to telecommunications providers, as they can divest from the infrastructure business through hardware-as-a-service from the cloud providers and enjoy OpEx for their own computing and storage requirements. It also presents an immediate benefit to the telecommunications and cloud provider joint customers, since every 5G site across the globe can now provide local computing, storage and container orchestration capabilities, as well as software-defined networking, using the widely adopted standard infrastructure as code (IaC) / configuration as code (CaC) of the hyperscale cloud providers. This benefit lowers the incremental cost to maintain solutions, creates resilient solutions that transcend telecommunications providers and unlocks previously unimagined use cases by simultaneously providing the infinite elasticity of the WAN cloud and real-time performance of the LAN cloud.
Reach for the Stars
But what about where 5G can’t reach? In the end, because money does make the world go around, there are still going to remain environments where 5G may not be the most economical solution. It is in these areas that Low Earth Orbit (LEO) satellite constellations are emerging again due to technical advancements beyond the limitations that saw LEO efforts, such as Globalstar and Iridium in the 1990s, fail. Starlink has shown impressive capability in providing broadband WAN access to the historically underserved / unserved areas of the world, and improvements are on the way! For example, laser-based communications between satellites can turn the constellation into a global service mesh, effectively mitigating LEO coverage window challenges previously only possible through Geostationary Equatorial Orbit (GEO) satellites. Cost and sustainability, however, continue to be a challenge for LEO constellations that drive up consumer costs. Amazon’s Project Kuiper presents an interesting approach to those cost challenges, natively integrating with AWS Ground Station as a Service and embedding edge cloud computing infrastructure within its satellites.
Between 5G ubiquity, hyperscale cloud providers providing embeddable, smaller form factors and the implementation of LEO constellations, one thing is for sure: the line that once defined a clear, clean “edge” is growing increasingly cloudy and will result in the transformation of digital experiences for all.