April/May/June 2020 I 27
As with many buzzwords, there are many
different definitions from many different
people and organizations, most of which
are correct in some way, but the unknown
cannot be precisely defined (Figure 1).
While all the aforementioned buzzwords
are important in their own right, they are also
driving the necessity for one in particular, namely
the edge data center (EDC). Because it can be difficult
to define individual unknowns, what happens when the
definition of one technology buzzword is driven by
undefined definitions of others? This is the case when
trying to define an EDC. Trying to define it is similar
to the response one receives when asking Siri what her
favorite color is. Never asked? Well, her response is “sort
of greenish but with more dimension.”
In the Telecom Industry Association’s (TIA’s) TIA
Position Paper, Edge Data Centers (2018), the many authors
and contributors from major ICT companies who comprise
IoT Latency
5G Edge
the TIA’s Edge Data Center Working Group ask the
question, “What is an Edge Data Center?” The answer
is as follows:
companies engaged in the development or installation
of EDCs, carefully research and assess EDCs as they
evolve over time.
The Uptime Institute, unlike TIA and BICSI, is a forprofit
data center entity creating its own proprietary standards.
According to its CTO, Chris Brown, “An edge data
center is a collection of IT assets that has been moved
closer to the end user that is ultimately served from
a large data center somewhere.”1
According to an EDC manufacturer addressing growing
latency needs, “We’re now seeing regional hyperscale
nodes of 20 to 60 megawatts in Ashburn and Chicago.
Even if it’s a 60 megawatt cloud data center, if it’s serving
that content locally, it’s an edge data center.”2
For data center designers and installers determined
to gain a better understanding of EDCs amid the Siri-like
definitions, two pieces of information are important:
• EDCs can be located within a conventional
ANSI/TIA-942-B data center.
• The standards and the breadth of content for
traditional data centers should also be applied
to modular, containerized, edge and hyperscale
data centers as cited in ANSI/BICSI 002-2019, Data
Center Design and Implementation Best Practices.
EDCs share many attributes with larger,
more traditional DC facilities however they
are designed to support widely distributed
(often cloud based) services. The location
of EDCs is driven by different business
cases—application latency, network capacity/
cost are common drivers—otherwise most
services would simply follow the trend and
end up in centralized facilities. EDC design
will require a new balance between redundancy
and availability. Operating many distributed
EDCs also implies a potentially large
impact in overall energy consumption.
The Siri-like definition provided by the TIA validates
the assertion that evolving technologies are difficult to
accurately define. Standards bodies, such as the TIA and
BICSI, are unaffected by the hype and the many often
conflicting reports attempting to forecast the future
growth of specific technologies including EDCs. Rather,
their work groups, consisting of data center experts from
HDC
Cloud
IIoT
Smart
Industry 4.0
FIGURE 1: Do we really know how the evolution of these
buzzwords will impact network requirements of tomorrow?