Understanding the Edge-to-Cloud Spectrum
Source: Linux Foundation Edge (LFEdge.com)
Just as the term ‘cloud’ was often thrown around needlessly and inaccurately in its early days, there is a fair amount of confusion surrounding the edge. What exactly is edge computing? Where does it start and end? How does it relate to other forms of cloud? The Linux Foundation’s Edge initiative has compiled an Open Glossary of Edge Computing to create a shared vocabulary. This blog post will build on its definitions to map the evolution and landscape of edge services as they empower enterprise applications and enrich end user experiences.
Edge Computing
Regions are retrenching after a period of rapid globalization. While the cloud originally centralized compute power, organizations today are beginning to distribute their applications into multi-cloud architectures that leverage on-premises, hyperscale, regional, and edge clouds. By processing data closer to the source, edge computing and 5G are helping organizations provide high-speed, low-latency services. From enabling fleets of self-driving vehicles to real-time health monitoring, the possibilities of edge are seemingly endless.
To take full advantage of the edge opportunity, service providers will have to deliver services across the edge-to-cloud spectrum. They will need to assemble unique combinations of value-added services to support their enterprise customer’s evolving application needs.
Centralized Data Centers
Centralized data centers refers to the hyperscale cloud infrastructure available from the likes of Amazon Web Services, Google Cloud Platform, and Microsoft Azure. Within this part of the spectrum, cloud providers leverage compute, storage and networks that can be consumed on-demand, with regions available around the world.
These centralized data centers are leveraged by a significant developer community that supports the majority of applications consumers use everyday (ie. Zoom, Slack, LinkedIn, etc). These application workloads typically do not require the low latency, real-time processing capabilities that are realized when running a workload at the edge.
The Service Provider Edge
The Service Provider Edge includes infrastructure and application platform services that are made available by telecommunications companies within their network edge (often in metropolitan areas). It brings together the internet connection accessed by devices with cloud capabilities delivered by service providers. In this way, it can be consumed alongside hyperscale clouds and may be included in hybrid- and multi-cloud architectures.
The service provider edge consists of the access edge and the regional edge - locations that are well suited for application workloads that may require lower latency, lower transport (egress) costs, higher performance, and/or compliance with regional data laws and restrictions.
Regional Edge
The regional edge refers to servers that allow different internet networks to connect and exchange traffic in regional data centers. It provides Internet Exchange (IX) points that allow users to access the internet. The regional edge helps enterprises minimize the ‘latency and network hops that would otherwise be required to reach a centralized data center, but which do not require the ultra-low latencies available at the access edge or the user edge.’ These data centers are best suited to support edge workloads with latencies between 30ms and 100ms.
Access Edge
The access edge serves the same function as the regional edge but is farther away from the radio access network (RAN). This could include infrastructure deployed at cellular data sites. The access edge would be leveraged to support ultra-low-latency workloads that nonetheless require predictable connections.
User Edge
The user edge, also known as the device edge, describes edge resources adjacent to end-users within the last mile network (the final leg of the telecommunications networks that deliver services to customers). It includes gateways, servers, and IoT devices. While these workloads often operate alongside service provider edge workloads, they have even lower latencies and represent a more diverse mix of resources.
On-Prem Data Center Edge
This edge includes infrastructure operating from within offices, factories, and colocation spaces as privately-owned data centers. Compute power accessed here is not as scalable, but often leverages security, automation and operational tools similar to those used in the cloud.
Smart Device Edge
The smart device edge is often what people think of when they think of the edge. It includes mobile and IoT devices outside of data centers that are capable of supporting compute power in the devices themselves. Graphics Processing Units (GPUs) are increasingly being integrated into these devices to accelerate analytics and performance, but the size of these devices limits their performance compared to the service provider edge.
Constrained device edge
The constrained device edge is sometimes referred to as the “tiny edge”. Constrained devices are battery-powered and connected locally to gateway devices. These devices have minimal computing power, but can perform near-instantaneous, microscopic calculations that deliver and receive real-time information for applications.
The Edge-to-Cloud Journey
Information technology evolves quickly, and the full potential of edge to enable faster and more efficient computing has yet to be fully realized. This is good news for regional telcos and service providers that can leverage their network and data center assets to deliver edge computing to enterprises. It is also good news for enterprises, as they will be able to pair their applications with cloud and edge services that match the speed, latency, and power they require.
Delivering, operating and monetizing edge-to-cloud services is a challenge for telcos. Whether your organization is just starting their journey, or your team is well underway and looking for a partner who can help you “git” there faster, let’s connect and explore how we can help your organization succeed in owning your destiny in the cloud.