Author: David McCall, QTS
Have you ever told someone that you’re in the process of moving your family into a new home? Chances are, you received a rapid mix of responses ranging from sympathy (“all the boxes…”) to envy (“you found the perfect spot!”) to curiosity (“how much did you pay?”)
In the data center industry, these kinds of conversations take place frequently and, just like with your new family home, there’s dozens of elements that contribute to the wheres, whys and hows of a Big Move. Buying an existing site (brownfield), renting, building greenfield...whether you’re talking a dream home or a data center, there are plenty of opportunities and an equal number of potential pitfalls.
When it comes to site selection, organizations seeking to launch or expand their data center operations have a number of factors to consider, not least of which is the question of whether to build a data center from the ground up or lease an existing facility. Those that are building will set very different priorities from those that are leasing, though there are a number of critical considerations that are universal to any site selection process.
For instance, all data center site selection projects should consider:
Location: Neither an engineer nor a real estate broker is needed to make this point: location is everything. The data center needs to be in proximity to the eyeballs it is intended for.
Economics: The cost considerations for a data center project start with competitive power and cooling costs, everybody gets that. Beyond those factors, however, business compatibility always plays a major role in the site selection process and data center operators wisely gravitate towards the most business-friendly environments. This typically comes in the form of tax incentives or tax credits which are frequently more valuable than the cost of the site or the monthly rent. This, in turn, can drive companies to invest in larger projects involving more hardware (such as hyperscale.)
Some sites may also provide a tenant improvement allowance to make it easier and affordable for companies to move in, make changes and expand.
Visibility and control: These aren’t two wholly separate concepts but rather they are two sides of the same coin. Visibility of your environment is required to make informed decisions. Accurate data is required for analytics to create business models. Control is required to act on the models in order to achieve business outcomes, manage risk, and react to threats. The ideal site should have a platform that pulls all of these business, operation, and critical infrastructure environments together under a single pane of glass or, alternatively, publishes APIs that enable businesses to integrate with their own platforms.
Redundancy: When considering the level of redundancy required of a new environment, there’s a few questions to ask: is this one of several in a metro ring or is it an active/passive site? Are applications part of a highly connected “mesh,” so that one site going offline is marginal in its impact to the whole? Those answers will determine how important redundancy is in the design and location selection.
Obviously key infrastructure components like power, cooling, and fiber entrances and paths are critical. To this list, add operations staffing, i.e. what happens if you have to evacuate the building? Is it operated remotely and, if so, where and how? What about all of the tools used to operate the facility; are those redundant? And while it typically holds true that the less redundancy, the less expensive the data center - in the case of some hyperscale or mega data centers - redundancy may actually result in cost savings.
Risk Assessment: Key to any business venture is solid risk assessment and that is especially true for a data center project. Weather and geography are clear risk factors that typically keep data centers from springing up along fault lines or flood plains, but additional considerations include: environmental risks (nearby habitats, protected land, etc.), the potential for physical and/or cybersecurity breaches, and risks to the power grid and other utilities, including fuel and water.
Compliance & Regulation: There isn’t less regulation in the world: PCI, HIPAA, SOC1, SOC 2, FISMA, FEDRAMP, HITRUST, and the list goes on. The point of compliance is to not only demonstrate that controls are in place, but also to regularly test against those controls. Partner with organizations that not only can meet the various compliant standards, but also have audit and compliance teams that can ensure those standards for the long-term.
Physical security: If an organization is going to have assets within a data center that are valuable, it will need physical security. Look for sites and designs that have a single point of entry, are hardened facilities and fencing around critical external infrastructure. CCTV around all ingress, egress, and public areas that not only records but also archives the recordings for future call back. Is the security staffing 24x7 or something less? Ensure the provider has a documented process for background checks of its personnel, especially its security force.
If building a data center from the ground up, all of the factors above are critical, but the focus should also include:
Facility Design: The fundamental facility needs of data center operators tend to boil down to two things: space and power. But beyond that, there are dozens of facility design elements that can have significant impact on the success of a project. Many of them address the core question: “does the space have the agility for the data center to grow, incrementally, without disrupting operations?”
Other basic design considerations are equally important, such as the number of loading docks, easy access to restrooms, and the amount of available office space. These provide answers to the question “does the space allow employees to come, go, and operate with ease?”
Set your priorities straight
Organizations seeking to lease data center space would shift to prioritize the following:
Expansion at the Speed of Business: A data center is only as valuable as the inventory it has available. The facility must have the ability to expand at the speed of business. Partners need to demonstrate the ability to provide all necessary inventory at the speeds their customers require.
Connectivity: While Hyperscale operators or mega data center owners (think Facebook) will likely connect to their own preferred network service providers, most enterprise data center users require access to multiple carrier options. Increasingly they will require SD-WAN capabilities, shoring up the connections between data centers and resolving connectivity issues.
Sustainability: Every data center developer will ask or will be asked the question “how green is it?” For most, sustainability is the byproduct of good citizenship and good economics with the mutual goal to reduce power consumption as much as possible. Developers need to consider a site’s designed power usage effectiveness (PUE) and improvements that can be made over time. In addition, with liquid cooling becoming a more standard practice, water conservation is quickly becoming a priority for operators. Developers shouldn’t consider sites that are not, at the very least, on a pathway to green.
Contract Expertise: After choosing a data center partner, companies want to be able to work out a contract that is equitable, properly structured and portable. By portable, we mean that companies are able to move services from one type of spend to another during the lifetime of the agreement. Site operators have an enormous advantage if they work with in-house counsel - attorneys that regularly work in the data center or IT arena - as opposed to external counsel. They understand the underlying technical jargon and they move quickly.
Management Team: Data center operators will inevitably work with their site’s operations team often enough that a solid partnership is bound to form. The site’s management team, however, may only be involved when a situation is at its stickiest. Operators need to be cognizant of just who it is that they are doing business: their reputations, their best practices, their knowledge of the industry. And most of all - are they good, reasonable people? The last thing a customer needs is to argue lease specifics three years into a five-year engagement.
Balance Sheet: Along with being trustworthy and knowledgeable partners, site operators should be able to demonstrate long-term financial stability. Particularly for hyperscale data center customers that are signing long-term leases of five, ten or 15 years, it’s imperative that they partner with an organization that is on strong financial ground and has the balance sheet to prove it.
Talent: If a business opens a data center, it stands to reason that they will want to employ a capable and dedicated staff to run it. Gifted employees are a rare and expensive commodity, so having a sense of a community’s talent pool can be a significant site selection factor.
24 X 7 Operations: If taking any services from a facility provider, ensure that it has a 24x7, fully-operational service center. Validate that the service center follows an international standard, such as ITIL, and will respond based upon severity and impact. The service center should be staffed by engineers to triage events, not pass them along.
This list includes just some of the considerations that go into a data center site selection, but every entity that enters into a project of this kind will have its own set of priorities and goals. Like a new home buyer trying to bring a sense of Feng Shui to their place, each organization must determine if a site will not just enable a data center to function, but if it will it enable it to function efficiently, thoroughly and in accordance with their needs.