A dedicated data center or
a funtional area within
A micro data center within
a building or tower shack
Mobile/Temporary In the neighborhood In the house On our person
TODAY TOMORROW BEYOND
Edge data centers defined more from their function of being close to the
data source than appearance or physical location. In the future, what we
consider an edge may be much different than we imagine today.
April/May/June 2020 I 35
tools, and remote functionality allow them to keep an eye
on everything, including infrastructure and components,
and to react or prevent issues as if physically there. For
availability, a commissioning plan done for the design of
a traditional data center should be done also for an MDC
with clear documentation and the operations and maintenance
(O&M) services agreement at hand. There should
be a plan to have available skilled technicians nearby on
a daily basis in order to ensure minimum downtime.
LATENCY, RELIABILITY, AND COST FACTORS
The “edge” is often associated with geography alone. While
that is part of the equation, it is usually the result of other
needs and requirements. Latency, cost, and reliability in
terms of quality of service (QoS) requirements need to be
considered when deciding how and when moving data
processing closer to the data source makes sense (Figure 7).
While building out a local edge costs money, sometimes
the cost is less than the hosting fees charged by larger
edges or even core data centers that may already be geographically
located close enough to accommodate latency
and QoS requirements. As the edge moves closer, there
may not always be a traditional conditioned space for the
data center to reside. Some EDCs may look a lot like and
be located in similar spaces and geographies as traditional
data centers. In other cases, an edge may be an isolated
part of a traditional data center that provides a specialized
service.
Contrary to some beliefs, EDCs can be multifunctional.
In the simplest configuration, they address a specific application,
function, or process, but they can also serve as edge
type colos serving many different applications and functions
of varying sources. As the number of devices increase
and latency requirements become more critical, costs and
reliability of transporting and processing data will become
areas of more concern and may force the edge to locations
perhaps unimaginable today (Figure 8). The industry has
yet to realize how 5G, IoT, and the other aforementioned
buzzwords will play out in their early stages of real-world
application. For now, determining where an edge is needed
and to be constructed should be a careful assessment and
mix of the latency, reliability and cost factors with a variable
weighting depending on the application.
FIGURE 7: Latency,
reliability, and cost are
three key factors to
balance when determining
the location and need
for EDCs.
FIGURE 8: As technology progresses, the look, size, and location of EDCs will adapt to meet the requirements.