If you’re like most people, you’re probably starting to hear more and more about edge computing. Maybe not a lot so far, since there are a lot of other things people are talking about right now like government shutdowns, “Why do they play full seasons in college football if we’re going to always wind up with Alabama and Clemson?” and the really big one, “If climate change is making the seas rise why is waterfront property still so expensive?” While all are interesting conversation starters, edge computing is going to prove to be a little more durable, so it makes sense to understand what it is and why it’s important
In the Beginning
Beginning in the mid-90’s a growing number of companies started outsourcing their data center requirements to third-party providers like Equinix, Digital Realty and QTS amongst others. To service, a customer base made up primarily of Fortune 1000 companies, these firms built facilities encompassing thousands of square feet in major cities like Dallas, Chicago, New York, and regions like the Bay area and Northern Virginia.
In recent years, the insatiable requirements of cloud providers like Google, AWS and Microsoft has catalyzed the “NFL-cities” model as they pre-lease entire data centers before the built. This “centralized” approach to locating data centers has been successful in supporting corporate applications, but the dynamics of information storage and processing needs continue to evolve, resulting in the emergence of edge computing delivered from edge data centers.
Driving the Move to the “Edge”
The need for immediate gratification
The acceptability of performance to end users is a function of latency (defined as the time required to complete a requested action and measured in milliseconds). The growth oasf new, and emerging, applications such as IoT, IIot and Virtual and Augmented Reality continue to grow within the business and consumer markets.
To an end user, latency is the reason that downloading a movie “takes so long,” but the number of milliseconds it takes to complete a function is measured in customer dissatisfaction and cost by content providers. Even at the speed of light the round trip from a central data center, a facility located in a major city, for example, can mean the accumulation of transmission costs. A study conducted by ACG Research estimated that caching content locally in a metro population can save approximately $110 million over five years.
IoT and IIoT Changes Everything
The Internet of Things (IoT) and its much bigger brother, (IIoT), are not some new technologies looming in the future-they’re here now. The success of corporate initiatives using these two modes of identification and tracking the real-time collection and accurate processing of data. Thus, the latency and reliability requirements of these applications will push the edge even closer to end users than in the commercial applications discussed above.
5G is a new standard for wireless communication that is being developed to support the elevated levels of requires high-speed bandwidth necessary to optimally support the voracious needs of applications like IoT, and big data as well as higher densities of mobile bandwidth users, device to device communications support and massive machine communications. To service these applications, 5G will require new network build-outs to enable the processing and storage of information at distances that are closer to end users than ever before.
What is an Edge Data Center?
In thinking about edge data centers, it is essential to understand that we are talking about a physical building and its geographic location. All edge-related decisions (how many are required, where should they be located) radiate from the purpose and user-base of the application(s) that will reside in the data center. With this in mind, an edge data centers can be expected to conform to the following criteria:
- Size: The proximity requirements for edge facilities dictate a smaller form factor than a conventional data center delivered at substantially lower cost (or lease rate).
- Built for reliability: Edge facilities feature designs that enable them to withstand severe weather and feature 2N, concurrently maintainable mechanical and electrical systems and backup generator sets.
- Geographically Independent: The defining question for a prospective edge data center customer is, “Can I put it exactly where I want it, like Cheboygan?” to which the answer should be “yes.” In other words, the edge should be wherever it needs to be.
- Easily Expandable: The volume of users and data shouldn’t be constrained based on the purpose of the facility. Edge facilities should be easily expandable.
Who Will Provide Edge Services?
For the foreseeable future, the avenues available to end users will vary based on their locations and requirements. From a national perspective it’s logical to assume that existing wireline and wireless (including cell tower providers) will become significant players in the edge marketplace. The maturity of their existing network structures and proximity to populations throughout the country provide them with a platform for collocating edge data centers that they can use to lease space to content delivery providers such as Akamai or even offer the same services to their customers directly. At present there are a small number of trials between edge data center providers and a few carriers are underway; with 2019 expected to see the number of tests, and their participants, increase substantially.
Since the ability to obtain edge support from one or more wireless/wireline carriers is a few years away until then companies with the need to support things like their IoT-based inventory systems across multiple factory locations can be expected to implement their own edge “networks” by collocating edge data centers at required locations.
Technology marches inexorably forward. The more information generated the greater the need to act upon it with a degree of rapidity that is unparalleled today. In light of this continued progression, the concept of edge computing will become essential with the fundamental question that will drive the development of increasingly smaller data centers being: “How much closer to the end user will it enable to us get?”