It is important to keep in mind the TIA standards and
BICSI best practices for traditional data centers throughout
the further exploration herein of EDCs, especially in light
of AFCOM’s dispute with research firms, such as Gartner.3
Gartner proclaims that the traditional data center is dead
and that “by 2025, 80% of enterprises will have shut down
their traditional data center, versus 10% today.”4 Clearly,
differing opinions and a host of varying predictions can
perplex even the best data center designers, installers,
and data center operators.
Determining what drives the need for an EDC and
why the enterprise customer may need one is a good
start to a better understanding of EDCs while mitigating
some perplexities.
DATA, DEVICES, IOT
AND EMERGING TECHNOLOGIES
Einstein said, “Not everything that counts can be counted,
and not everything that can be counted counts.” Many
people have come to accept this, but can the ICT industry
expect this to still be true as technology evolves? Various
market projections forecast that there will be between 25
and 50 billion connected devices by 2022, creating mass
amounts of data to be processed, transported, and stored.
Devices today already collect more information than most
enterprise managers and data center operators know how
to use. However, the devices that collect the data do. The
amount of data, the amount of devices and the amount
of data that the devices collect continues to increase
at a rapid rate.
They collect data that data center operators never
thought needed collecting. Some of these devices act
as peripheral devices with IoT that have one smart
device feeding or monitoring many other devices
28 I ICT TODAY
that communicate or process data essentially as a sub-
network. Advancements in sensors, artificial intelligence
(AI), machine learning (ML), virtual reality (VR), augmented
reality (AR), information awareness (e.g., health
vitals), and other content enable the collection of endless
amounts of data. Enterprise managers are finding
ways to utilize “newly able to be collected” data to predict
required responses or to drive other tasks. Soon,
everything that counts will be able to be counted, and
everything that can be counted will eventually count.
DEFINING LATENCY
All data may be important, but do they each require
different response times to process and enable the next
action? Some data has an expiration date. With so much
continuous data being collected, current data expires
almost instantaneously. This drives the need for ultrafast
reaction time for some process requests. This reaction
time is dependent on another aforementioned buzzword,
latency. Dictionary.com defines latency as “The time
required online or in a network for the one-way or roundtrip
transfer of data between two nodes.” Verizon defines
it slightly differently as “The time it takes for data to travel
from one user, over the network to the central processor
and back again.” The mere mention of latency drives
people to want the lowest possible latency, but is it always
needed, practical, or affordable for the given task? Many
in the ICT industry often hear that 5G will be the solution
for networking problems.
While there will be increased bandwidth and speeds
with 5G, they are not directly related to improved latency
as many assume. In fact, even with 5G, latency will
become the new bottleneck to the development of
apps and other process functions. Many people in the
ICT industry believe that there will be more and more
EDCs located at the base of cellular towers and other
5G aggregation points.
Crown Castle’s CEO, Jay Brown, tells another story, “At
this point, I really don’t see data centers playing a significant
role in our long-term strategy. We think the opportunity
for us really relies around towers and then the use of
fiber for small cells.” Crown Castle is the largest cell tower
owner and developer, owning and operating approximately
40,000 cell towers in North America and providing
While there will be increased
bandwidth and speeds with 5G,
they are not directly related
to improved latency
as many assume.
/Dictionary.com