Insight and analysis on the data center space from industry thought leaders.
How Mainstream 5G Will Disrupt the Data Center Status Quo
The US networking infrastructure continues to lag behind developments on the mobile side, resulting in a system ill-equipped to handle the demands of 5G communication.
February 14, 2019
Albert Ahdoot is Director of Business Development at Colocation America.
In a January 2017 report, IHS Markit positioned 5G as “a catalyst that will thrust mobile technology into the exclusive realm of GPTs (General Purpose Technologies).” Already populated by the likes of the printing press, the steam engine, the automobile, and the internet, this realm encompasses a wide range of revolutionary technologies that “redefine(d) work processes and rewr(ote) the rules of competitive economic advantage.”
If the recent 2019 Consumer Electronics Show (CES) is any indication, many major telecommunications players are similarly bullish on 5G’s potential as a competitive differentiator. Over the course of the four-day event, Samsung outlined its plans for an end-to-end 5G connectivity solution, AT&T touted its 5G mobile network that went live in 12 pilot cities in December, and Sprint revealed it had completed the world’s first 5G data call across the 2.5GHz spectrum on a live commercial network.
This widespread buy-in is an important first step toward actualizing the $3.5 trillion of annual output IHS Markit expects the global 5G value chain to generate as soon as 2035, but there remain a number of formidable obstacles standing between the current state of mobile networking and universal 5G connectivity. Perhaps most significantly, developments on the fixed side of the United States’ networking infrastructure continue to lag behind developments on the mobile side, resulting in a bifurcated system ill-equipped to handle the demands of 5G communication.
An Ongoing Technological Imbalance
Promising mobile internet speeds of over 10 gigabits per second—roughly 100 times faster than those delivered by 4G LTE—and end-to-end latency of under a millisecond, the crux of 5G’s potential rests with its unparalleled speed. In tandem with increased bandwidth—another noteworthy improvement 5G promises to provide—this speed will dramatically accelerate the maturation of technologies like autonomous vehicles (think both self-driving cars and commercial aerial drones), telemedicine, smart grids, and any other remotely operated, fault-intolerant system.
Because the systems that stand to benefit most from 5G are inherently fault-intolerant, it’s imperative that the country’s entire networking infrastructure is capable of supporting 10-gigabit speeds and sub-millisecond latency. Unfortunately, as things stand, few stakeholders have taken the initiative to prepare the fixed backend of the country’s networking infrastructure for 5G.
The overwhelming majority of developments showcased in forums like CES pertain to “last-mile connections” between network routers like cell towers and mobile devices like smartphones and Internet of Things (IoT) sensors. While connectivity over this last mile has been revolutionized by everything from beamforming to massive multi-input, multi-output arrays to millimeter waves, connectivity between networks’ fixed edges and their cores—i.e. their server infrastructure—has not meaningfully evolved since the introduction of 4G nearly a decade ago.
Against a backdrop of skyrocketing data production, this insufficient fixed infrastructure will leave companies angling to leverage the potential of 5G with a single surefire solution (at least for the time being): move their data processing operations closer to the network edge.
Solving the Latency Problem
Simple physics dictates that the less distance a data packet must traverse, the less latency an end user will experience. Currently, users’ requests are often shuttled hundreds or even thousands of miles between the nearest cell tower and a centralized data center, but the relatively slow internet speeds delivered by 4G mean that the latency resulting from this travel distance isn’t noticeable to the end user.
As 5G becomes increasingly mainstream, however, continued reliance on massive centralized data centers will largely offset 5G’s benefits by exposing end users to (relatively) high levels of distance-induced latency. Until the country’s fixed wireline infrastructure is substantially upgraded, the best—and, arguably, the only—way for companies to take advantage of everything 5G has to offer is to diffuse their data processing.
To realize the potential of 5G, companies must figure out ways to both rapidly process new data and move caches of existing data closer to end users. Achieving the former has been the impetus behind the recent rise of micro data centers, or “containerized data centers,” located right on network edges (i.e. at the base of a cell tower). These small-capacity servers are able to fulfill routine requests from end users without running the workloads through a traditional data center, all but eliminating latency.
But while micro data centers will play a pivotal role in, among other things, the evolution of the IoT, they’re incapable of fulfilling requests involving complex data processing or access to extensive cached information. To drag these kinds of workloads into the 5G era, companies have little choice but to geographically diversify their data center portfolios.
For enterprises, building multiple new data centers at strategically located points across the country might be within the realm of possibility, but for most small and mid-sized businesses, this is well beyond economic feasibility. As such, for these companies, data center colocation often represents the easiest way to expand their high-powered data processing footprint.
Colocation lets companies rent appropriately-sized server infrastructure in multiple facilities simultaneously, enabling them to process and store data in close proximity to their end users even if the users aren’t concentrated in a single geographic market. If, for instance, the majority of a company’s user base is located in Southern California but sizable minorities of the user base are located in the Tri-state area and the upper Midwest, the company could maintain its own primary data center in Los Angeles while renting a server rack or two in colocation facilities in New York City and Chicago.
The Way Forward
Ultimately, data center diversification—whether through micro data centers, increased investment in proprietary data centers, colocation, or, most likely, a combination of the three—will be a critical component of any successful 5G implementation. Instead of viewing this preparatory work as a prohibitive inconvenience, stakeholders should embrace it as a necessary step in enshrining mobile technology in the GPT pantheon.
Indeed, the emergence of almost every GPT has been prefaced by a great deal of foundation-laying. The full effect of the printing press didn’t take hold until mass literacy became a societal priority. The full effect of the automobile didn’t take hold until governments invested in sprawling networks of paved roads. The full effect of the internet didn’t take hold until personal computers became affordable to the common consumer. And the full effect of mobile technology won’t take hold until stakeholders across the telecommunications spectrum adequately prepare their mobile and fixed infrastructures for 5G connectivity.
Opinions expressed in the article above do not necessarily reflect the opinions of Data Center Knowledge and Informa.
Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating.
About the Author
You May Also Like