With the adoption of AI, IoT, 5G and other applications on the uptake, enterprise are finding themselves having more data and more data types to process and store. Consistently heavy workloads can strain traditional on-premises DCs. While more compute, storage and other assets can be added to such facilities, there are limitations as to how much more space, power supply and cooling can be made available. What’s more, such add-ons can be technically complex and there is also a risk of day-to-day operations being disrupted.

To get around such limitations, more and more enterprises are complementing their on-premises DCs with a constellation of remote ones. While the capabilities that these smaller facilities provide are available from cloud providers, this option is not open to enterprises that operate in countries with exacting data sovereignty regulations.

A Matter of Choice
Others that choose to operate their own remote DCs may also desire to maintain autonomy over data and minimize their dependency on external providers. Moreover, remote DCs engender multiple-redundancy data backup while trimming data transfer costs and improving latency and security. Another advantage: they can be pre-planned and -sized for future upscaling.

Remote DCs aren’t always located in another country or continent; they could just be in the next state or region or even the other side of a city. Size-wise, they aren’t as big as traditional DCs but occupy more space than edge DCs.

Key Factors
So what are the key factors to consider when planning new builds of remote DCs? Start with these:

Manageable High-Performing, High-Density Infrastructure
Populating remote DCs with as many high-performance assets as possible makes the most use of their modest size. Each of these assets require a power supply so PDUs that feature a high density of power outlets are the best choice. The Server Technology PRO4X and Raritan PX4 intelligent rack PDUs both feature high density outlets that combine an IEC C13 and C19 into one outlet, accommodating C14 and C20 plugs, with alternating outlet and branch power distribution in a single column of outlets down the entire length of the PDU.

The high-performance PDU also measure total harmonic distortion at the cabinet and device levels, offering operators the most complete and accurate set of internalized power quality monitoring and metrics. For ease of management, it is best to cascade PDUs in a daisy chain through either USB or Ethernet ports. Such network connectivity makes it easy for operators to access, administer and upgrade each device in the chain over the Internet from anywhere.

Right-Sized PDUs
PDUs play a vital role in DCs, allocating the right amounts of voltage and current to servers, networking equipment and storage hardware at the right times. As with other assets, rightsizing PDUs and developing the ability to accurately assess capacity needs are critical.

Start by determining what devices will be deployed in each rack. This will determine the PDU outlet type(s), number of outlets, and the types of plugs, and help teams estimate the kW/kVA needed for that rack.

Next, use the nameplate power rating of each device to calculate the power requirement. This is achieved by adding up these values and multiplying by 70%. Keep in mind potential growth for additional headroom.

Choosing the right PDU is key to ensure that your critical IT equipment will be powered reliably. Raritan, a brand of Legrand, offers a portfolio of 600 intelligent PDU base models with more than 1,200 SKU variations.

Remote Serviceability
Remote serviceability of remote DCs is vital as they typically lack on-site technical staff. Historically, most enterprises have only thought about remote serviceability in terms of basic management capabilities such as system status alerts, soft reboots, and software patching. But as DC infrastructure increases in density, sophistication, and criticality to the business, remote serviceability and management capabilities must increase as well. These may include full remote KVM sessions, hard reboot, temperature threshold alerting, and more.

Raritan offers solutions that provide BIOS-level remote management regardless of where the equipment is located. They include KVM-over-IP switches, serial console server/terminal server, serial access modules, and CommandCenter secure gateway. All are based on the Xerus hardware and software technology platform. Using them, DC operators can successfully install, configure, update and troubleshoot remote computers, networking and facilities.

Next Month: We look at other key factors for remote DCs. Stay tuned!