Subscribe
Close
  • Free for qualified executives and consultants to industry

  • Receive quarterly issues of Area Development Magazine and special market report and directory issues

Renew

Edge Planning Considerations in Data Center Development

Low-latency requirements, which surpass the capabilities of centralized cloud-delivered services and data centers, have given rise to the need for smaller data centers, deployed at the edge of their networks.

Data Centers 2019
In a world where an estimated 6.1 billion smartphones will be in operation by 2020, traditional paradigms for how data is processed and stored have given way to a variety of prospective data center alternatives available to businesses. This doesn’t mean that dedicated and collocated data center options don’t, and won’t, continue to exist, but it does mean that they’ve been augmented by a host of new prospective solutions with new arrows regularly being added to the quiver.

New technologies and capabilities like IIoT, IoT, Augmented and Virtual Reality (AR/VR) are providing an abundance of opportunity for data center operators and providers to undertake a variety of activities including improving their internal operations, enhancing customer experiences, and delivering new products and services to their customer bases. The common thread that defines this evolution is the increased degree of time sensitivity, referred to as latency, associated with the transmission and processing of information to perform the necessary operations.

These low-latency requirements surpass the capabilities of centralized cloud-delivered services and data centers. Thus, organizations planning to implement applications ranging from sensor-based tracking and purchasing systems, to enhanced diagnostic ability via instantaneous patient record or radiological image access, to enabling customers to download video content with no perceived delay need to place compute and storage functionality as close to the customer as possible. The combination of these elements has given birth to the need for smaller data centers, deployed at the edge of their networks to support these new low-latency requirements. However, just like the process of implementing any data center solution includes questions that must be answered, edge data centers are no different. For example, seemingly trivial matters like, “Where am I going to put these things?” must be addressed. Since the devil is always in the details, the key element of successful edge implementation is, of course, planning.

Catching the Wave: Proceed with Caution
Edge data centers are a vital element in the coming “wave of the future” regarding solutions that will bring things like content and processing power continually closer to end-users. Naturally, the common reaction tends to be, “I’ve got to get me some of those.” However, the real data center and development professional knows that haste may not only make waste but leads to undesirable consequences as well, so proper planning is essential before the first PO (purchase order) is even a gleam in IT’s eye.

EDGE locations can be spread over a broad area or within a narrowly defined region, but in each case, they will be physically separate from each other. Elements of the Planning Process
The process for determining edge data center requirements, configurations, and geographic locations relies on both business and technical evaluations.

Business Considerations — Establishing a network of edge facilities is a little more involved than purchasing a few units advertised as “edge data centers” and locating them in the hinterlands. The business phase of these initiatives is devoted to examining the project based on existing IT structures and operations to define its scope and provide the objective data required to translate the business requirements into the physical template(s) to effectuate. The answers to three basic questions facilitate the business portion of your edge planning:

1. What is the current data center structure, and which applications reside in each facility?
While evaluating the requirements for an edge data center project may seem superfluous, since it’s already known that the underlying requirement is to put specific functionality closer to end-users, virtually all edge initiatives will be contingent on the existing network and, oftentimes, existing applications. Based on the inherent inter-relationship between the existing core and its desired augmented edge, existing and proposed applications must be reviewed to determine things such as their latency and security requirements. In other words, what is the optimal as well as maximum level of acceptable latency?

2. Who are the “end-users” (internal and external) to be supported and what are their specific requirements (i.e., what is required to provide the necessary level of service?)
Essentially, this effort is designed to develop specific profiles of the “end-users” of the edge-centric application(s) and establish the satisfaction criteria for each group. For example, an initiative to provide medical personnel with faster access to patient records and radiological information will be used differently by a physician whose primary requirement may be speed and administrative staff whose focus is on ease of usability and information security.

3. Where are prospective “end-users” located, and how many reside in each geographic area?
This analysis isn’t conducted solely to identify new edge locations but also to determine the required configurations for each edge data center. Depending on a variety of factors, the number of applications to be supported and user volume, for example, the quantity of hardware such as racks and servers may vary between locations, thereby requiring more than a single edge data center.

The answers to these geographic questions have a substantial impact on implementation planning. Physically, edge data centers do not fit neatly within many municipalities’ code structures. As a result, additional time should be built into implementation plans to compensate for potential issues regarding permitting requirements, zoning, and even municipal jurisdiction disputes until precedents are set to address a rapidly growing number of edge installations.

Edge data centers are a vital element in the coming “wave of the future” regarding solutions that will bring things like content and processing power continually closer to end-users. Technical Considerations — Just as form follows function, the business analysis findings must be translated into physical requirements. The specifications documented in this phase serve as the basis for multiple activities including vendor review and selection, identifying the potential need for site acquisition and development support, and the physical requirements for each site. The process of making these determinations focuses on developing answers to the following questions:

1. Based on the business analysis, what are the configuration requirements (racks and cabinets) for each desired location?
Each site should be looked at individually to determine its initial, and potentially future, hardware, power, and cabling needs to accelerate the delivery process and clarify site requirements to simplify data center installation.

2. What are the physical requirements (fiber, power, etc.) for each location and are there known existing sites that meet these criteria and/or will additional locations need to be identified and acquired?
Indeed, the ability to locate edge facilities on existing corporate property is desirable, but due to the inherent remote nature of an “edge” site relative to both existing corporate locations and proximity to end-users, the site selection process for edge data centers will be similar to that of larger core facilities. As a result, access to fiber and power, site availability, and the need for site development will continue to be primary evaluative criteria.

3. How will you implement edge units at existing corporate locations versus newly purchased sites?
Edge implementations can typically be expected to take place in multiple locations within an abbreviated time, and requirements may vary between a corporate-owned site and “greenfield” sites. Processes and procedures may vary by individual location.

4. Who will perform the necessary pre-installation and installation activities at each location?
Since pre-installation and installation support varies between equipment providers and/or end-users, specific entities need to be identified who will be responsible for:
  • Obtaining all required site development and building permits
  • Coordinating all related scheduling
  • Interfacing with utility and fiber providers
  • Supervising the physical installation of the facility
  • Determining if these responsibilities differ between sites
5. What is the support plan for these new implementations?
Edge locations can be spread over a broad area or within a narrowly defined region, but in each case, they will be physically separate from each other. These geographic considerations mean that your support plan needs to determine who — the end-user or the data center provider — will monitor your sites and perform routine maintenance, emergency support, and spare parts availability.

In Sum
Implementing data centers has always been a detailed and complex exercise. Edge data centers will have a multiplicative effect on these processes. Operating and monitoring remote data centers necessitates that organizations take a hard look at how these network extensions will impact current modes of operation. Edge data centers will provide a powerful tool for the implementation of emerging applications, enhancing the end-user experience and increasing customer satisfaction, but will also present new sets of challenges that require careful analysis and planning.

Exclusive Research