As many of my colleagues and counterparts will agree, a digital transformation discussion typically focuses on a couple of key points that are in fact very germane to the discussion – WAN and cloud. These two subjects seem to manifest the most emotional attachment from to Solution Architects when engaged in this discussion and rightfully so. One simply cannot exist without the other and both should rightfully receive equal airplay in any digital transformation conversation. But as broader architectural constructs, they are both ultimately designed and leveraged for a greater purpose.
The underlying topic that arguably deserves far more attention than is typically afforded to it has far more to do with an organization’s bottom line or ability to execute the mission. Applications are critical to an organization’s success and should always remain central to the digital transformation discussion and any eventual end game. We know that most organizations are challenged with determining which applications are best suited for the cloud, as opposed to those that will likely not make the cut, eventually meeting an unceremonious end through attrition. We also know that with an ongoing gravitation to more efficient “regionally distributed architectures,” centralized application models of old are giving way to a more regionally distributed model that assures optimal “geo-strategic alignment of applications, data and eyeballs.”
Observations from the field
An active customer of mine, who I consider to be an advanced practitioner of digital transformation as it relates to developing a practical roadmap from start to finish, offered an interesting take on the process. Simply put, he surmised that in all likelihood his extensive suite of applications all deemed mission critical in varying measures to a large government agency (which will remain unnamed) would ultimately be relegated to a 70-20-10 rule. He predicted that roughly 10% of existing applications would move to the cloud in their existing form or in some virtualized variation, while another 20% will require various levels of finesse, significant re-architecture, etc. He’s also conceded that nearly 70% (legacy, proprietary, CAPEX-dependent applications) would simply not make the trip and would be phased out via attrition. Based on recent industry observations, these all seem like fairly reasonable assumptions. More often than not, the adoption of OPEX-centric cloud native alternatives makes more sense from an ROI/TCO and performance perspective. His particular transformation journey is still in play, so the jury’s still out, but I look forward to a follow up at some point to see how things ultimately pan out for him and his organization.
What we’ve definitely learned in our observations with partners and customers here at Equinix is that a multitude of new use cases continue to emerge as a result of emphasis on a regionally distributed vs. centralized approach. Most have a focus with distributing an application with some element of private, public and hybrid clouds across a number of metros within a given region that solves for speed of light issues and enables the application to be more responsive, significantly increasing the quality of experience (QoE) for customers/users of the application(s). Streaming a centralized application to the same regionally distributed user population is simply untenable at this point. The multiplexing of varied content from disparate sources that represent an application stream (e.g., biometrics, distributed analytics, etc.) requires edge-based proximal adjacency to core services (clouds/networks) and geo-alignment with those who consume it. Consistently converting data into real-time actionable information simply cannot be done otherwise.
An application-centric architecture
This is where things really get interesting, particularly as it relates to the earlier discussion regarding a WAN and cloud architectural construct, which has become far more attentive to QoE. Far surpassing the “best effort” approach of largely static and costly network routing of the past – containerized, platform-agnostic applications are now distributed at the edge. New “intelligent” networks that are essential in accessing and distributing these applications are now more application-centric with policy-based routing (SD-WAN) that employ advanced metrics that can be configured to dynamically choose the best path for all layers of a given application.
Complimentary technologies like Network Function Virtualization (NFV) that decouple traditional network devices from physical boxes and virtualize them, i.e. firewalls, load balancers, intrusion detection devices, WAN accelerators, etc. provide a platform more conducive to dynamically consumable services. A key building block of Software-Define Networking (SDN), distributed NFV enables the positioning of a regionally-distributed architecture – a key facet of application optimization. A more flexible, application-centric and geo-strategically located transport and security stack enables us to meet even the most demanding application performance requirements – both when and where they are needed most.
The ultimate QoE enabler
Now more than ever, applications are expected to continually evolve (DevOps), always be available and provide accurate, real-time information to consumers. And, the stakes rise significantly when those applications become integral to, for example our national security. Border protection, travel safety, and infrastructure management and protection are just a few examples where critical functionality is dependent on application optimization.
Prolific consumers of information have risen societal expectations, at least in part, because of the richness and broad availability of information through the applications they use, particularly from mobile platforms. Initially developed and operated largely in isolation from a central location that often limit their reach and performance to a growing disparate and distributed user population – regional distribution, intelligent integration and geo-strategic alignment of applications and data with their network, cloud and consumer counterparts have become the ultimate QoE enabler.
Check out the Interconnection Oriented Architecture™ (IOA™) Blueprint on deploying applications at the digital edge.