We at Equinix — the world’s largest interconnection platform provider — have made some key observations over the past several years, particularly relating to the evolving strategy of deploying wide area network infrastructures. The following provides a narrative for the illustration below, with emphasis on two sharply different methodologies in play from both the past and present.
The traditional approach for designing and deploying a WAN architecture “Before Interconnection,” as illustrated in the diagram below on the left, is usually centered around a single centralized hub or data center where an organization’s applications and data are hosted. As was usually the case, an MPLS network was deployed with its core nestled within the centralized data center while extending hundreds if not thousands of MPLS “tentacles” from the core to far-reaching locations across a given region or, in some cases, globally. As the limitations of the speed of light will often dictate, this latency-prone approach at scale predictably provides varying levels of quality of service and quality of experience for often widely disparate user communities collectively with varying proximity to the network core. In retrospect, this approach is viewed by many as a network-centric architecture.
With the proliferation of geo-distributed cloud consumption, next-gen applications and a far higher QoE expectation from users, many enterprises, including federal government organizations, are now viewing a new application-centric or “regionally distributed architecture” as a far more viable alternative to its predecessor. Replacing MPLS sprawl with purpose-built regional hubs, such as those illustrated in the “With Interconnection” diagram below, that more effectively align applications, data and eyeballs within metros across a region has proven far more effective from both a cost and performance perspective.
This approach is further enabled with emerging application-centric networking components such as SD-WAN, which enables an organization to more intuitively leverage the largely static WAN resources of the past with new policy-based routing that enables route and usage prioritization based on the needs of the application and its consumers. As an extension of that architecture, Equinix customers can also leverage our SDN-enabled Cloud Exchange Fabric™ (ECX Fabric™) platform that, by design, provides customers with private, high-speed proximal adjacency to clouds, while also enabling dynamic intra-Equinix, inter-metro/inter-region interconnection provisioning by the customer when and where it is needed. For example, a customer who occupies two locations at Equinix can dynamically provision up to a 10G circuit between their Atlanta and Seattle locations within 15 minutes. They can then opt to keep that connection in place for a day, a week, a month or indefinitely — it’s entirely based on their requirements.
The elastic application-centric networks that were being conceptualized just a short time ago have finally begun to arrive, and the results have been truly game-changing. Both cloud and WAN transformation have become integral to most digital transformation discussions where intelligent transport and the convergence of legacy and cloud environments (hybrid and multicloud) have become more of a mainstream consideration moving forward.
Equally important, many federal government customers view this new approach as a means of simplifying their respective cybersecurity schema. Geo-strategically instantiating, regionally distributed hubs enable organizations to effectively shrink their attack surface with fewer points of ingress/egress from within trusted enclaves — these new hubs are each situated behind prescribed, agency-specific security stacks that effectively harden them from intrusion. As government entities continue to employ new cutting-edge technologies (biometrics, facial recognition, etc.) in the interest of national security, they’re also finding the benefit of cloud-adjacent, regionally distributed architectures to effectively deliver those applications with the intended results.
In short, the emphasis has changed from a network- to a decidedly application-centric architecture, assuring far superior performance and cost avoidance. Moving applications, data and network distribution to the edge in a regionally distributed architecture should be a key consideration as part of an organization’s digital transformation strategy. Our customers have discovered the distinct advantages of doing so while leveraging an ecosystem of 500+ cloud and 1,800+ network providers that operate within our carrier-neutral facilities across the globe.
You may also be interested in reading: