Managing Data Centers at the Edge

15. February 2022

5 Min read

Contents

Written by Carsten Ludwig

Market Manager Data Center

Many interests and …

The most important task of edge data centers will be to provide computing performance on the periphery of the cloud and wide area networks, to relieve the network, and to offer better performance. They are as close as possible to the point of action, where data has to be aggregated, analyzed, or processed in real-time.

This is where all conceivable interests meet directly. Some users only need a rack or even share it with others. Larger customers expect independent cloud services or alternative colocation packages.

Maybe a customer is planning a modest 40 Gigabit Ethernet. At the same time, another expects the host to pave the way to 400 Gigabits during operation. Others are looking for an operational base for 5G applications or resources for load-sharing as well as additional capacity for new AI and streaming services.

… many functions in a confined space

The infrastructure and technology of edge data centers correspond to the logic and functionality of commercial cloud data centers. But they are small, very small.

Ultra-High Density (UHD) is all-important for racks and patch panels. More than 100 ports per height unit and hybrid capabilities would be desirable. The patch panels should be able to facilitate a mix of media and applications. All of this requires a certain degree of sensitivity when it comes to assembly, handling the connectors, and cable management.

Edge data centers with fiber optic hyperconnectivity in every direction: to the cloud, to cellular phone networks, to neighboring edges, and to users. In other words, there must be space in the floor and in the cable ducts for high count fiber cables.

A challenge for design engineers …

Today, we can expect edge providers to operate dozens or even hundreds of edge data centers concurrently. They are spread across many locations in urban and suburban environments.

These can sometimes be uncomfortable locations outdoors. In any case, edge data centers and their casings should be exceptionally robust and also secure. In a way, the protective walls of a large data center are compressed. Miniature data centers must function largely autonomously without needing maintenance.

Standardized formats make scaling, stacking, and linking easier. Pre-terminated racks, one-housing systems, and turnkey containers can be set up quickly and maintained efficiently. The plug-and-play principle applies to connectivity and IT. Cabling and connectivity technology should be self-explanatory and easy to maintain. All this poses a challenge for planners and design engineers.

Furthermore, the small space needs to be supplied with power and UPS, fire protection, air conditioning, cooling systems, sound protection, dust protection, lock, and access control – another design challenge.

… and for operators

It seems impossible to manage a large number of remote edge data centers in a traditional way. It would not be financially viable to keep expert personnel available at each site. There simply aren’t enough skilled workers.

The result is that edge data centers must be able to be managed centrally and remotely.

• The DCIM system must offer extensive control and intervention options similar to those of an in-house data center.
• It must be possible to display performance, assets, operating data, and security status on an intuitive-operation dashboard with just a few clicks of the mouse.
• Likewise, life cycle management, risk assessment, the planning of patching and maintenance tasks, and cost control should be immediately available.

A fully automated view of each individual link is also crucial. Operators need real-time digital information about the status of each patch panel as well as a complete digital profile of the network infrastructure. They want to be able to precisely locate performance losses and manipulations at remote premises in order to prevent failures, and, in the event of a loss of performance, to identify the causes immediately. This shortens the time it takes to repair the damage and optimizes performance for end-users.

If a technician has to work on-site, the operator should be able to monitor and manage the patching jobs centrally. Signaling directly at the ports can give precise work instructions.

One solution: R&MinteliPhy net and its hardware-based functions

Our software- and hardware-based management system addresses the challenges of EDGE and focuses on all relevant performance and security aspects of these new IT network elements:

Have you already tried inteliPhy net 3.0? Request a free trial version now. You will receive the installation instructions and a test license code for 30 days from us by e-mail.

Are you planning an edge data center right now? Then you should read my planning recommendations first.