By Tim Kittila, Director, Data Center Strategy, Parallel Technologies, tkittila@ptnet.com
As the consumption and the creation of data multiplies at exponential rates, so does the demand for data centers and colocation facilities. These data centers provide critical network connectivity points, provide hosting services for private and public cloud options, or host variations of enterprise data.
Some would argue that as a result of this current “outsource” swing in the market place, additional data centers are being added to accommodate the growth in colocation and cloud options. Others see shift in the market place as network demands are forcing our data to be close to the end-user, calling for the growth of new “edge” data centers. The various demands and growth of IoT (Internet of Things), data availability, data storage, data processing, etc., is pushing the consumption of data center space. As a result, we see new data centers being required as the commodity of “data” becomes more valuable than the commodity of oil.
However, data centers can’t just be built anywhere. It requires proper site selection to ensure reliable operations and profitability. Like any building project, data center site selection is a complex process of economical and geographical considerations as well as weighing needs against risks factors to find the best location. Selection should also account for factors such as connectivity and proximity. In each category, use long and short-term goals to determine what can be “given up” in order to balance a “need” or requirement. It is a delicate balance of maximizing connectivity and mitigating as much risk as possible.
Accepting and Mitigating Risk
Time and time again, we’ve seen in the news how disruptions to a data center can cripple business operations. Companies simply cannot afford to have a data center outage.
Risks can take many forms –from weather events to natural disasters and environmental hazards. Unfortunately, there is no perfect or safe place in the United States (or the world) completely void of risk. There are risks and drawbacks associated with every location – whether natural or man-made.
The conversation isn’t about how to avoid risk (as it is inevitable) but rather, it’s determining what risks are of grave concern and weighing it with what risks are acceptable based on company goals and objectives. Once that is determined, the next step is to determine how to mitigate against that particular set of risks.
For example, a client recently approached me about building a data center in a metropolitan area that was situated in a historic flood plain. Because this center was to be a colocation facility, it was important to the client that the location be highly visible and easy to access, but most importantly, that it was ideal for connectivity. In this situation, the client was willing to take on the risk (and added expense) in exchange for the connectivity, visibility and accessibility benefits. In order to mitigate the risk, it was decided to raise the building elevation.
Maximizing Connectivity
Hands down, the first and more important area to consider when determining a location is connectivity. Its’ the driving force behind all data center decisions and encompasses three elements: telecommunication/fiber optic infrastructure, power/utility and accessibility. Companies will often take on more risk for better connectivity because without it, data center operations aren’t able to deliver the value needed to meet business goals and objectives, not to mention keep customers happy.
Telecommunications Infrastructure
Since much of the power of the data center lies in its ability to provide connectivity to its internal and external customers – making telecommunication a critical component to selecting the right location. Even with miles of fiber optics being laid each day, access to the telecommunication infrastructure drives a lot of the constraints around data center location. In rural areas, for example, connectivity is still a concern.
When it comes to telecommunication infrastructure ask the following questions when evaluating this part of site selection:
- What kind of fiber is available? This will affect speed and transmission and will lead to a discussion about transaction time and latency – which are also important factors in the selection.
- What is the fiber backbone route and its proximity? This will help gauge addition investments required from backbone route to the exact data center location.
- What providers are in the area? This will identify the carriers in the vicinity and the type and level of support available.
- What is the fiber redundant pathways? Clients rarely will consider sites without seeing means and methods for a redundant pathway.
Power Infrastructure
Data centers require a great deal of power to run equipment. And since power is the primary cost associated with operating a data center, access to a reliable power source is a key element of the selection process.
Access to power involves two things: delivering power to equipment within the data center and getting power delivered to the data center itself. As power is evaluated, it is important to keep in mind that it not just about accessing power, but also the facility’s overall access to the power grid. Some questions to ask in this area include:
- What is the maturity of the grid? This will help determine the reliability of the grid not only today, but in the next ten years. It also comes into importance with the considerations for soft-loading systems and utility interface.
- Where are the power stations, substations and facility feeds located? Proximity to these key points of transmission will factor into reliability.
- Do we have access to multiple substations or multiple providers? The ability to access more than one power grid will enhance the viability of a particular location. These come at much greater costs, but knowing the availability to have this option, will leave you flexible for options in the future.
Another important consideration beyond choosing a location with access to a reliable power grid are utility costs. Compare the cost of electricity across the various options inconsideration. Another client example is that after reviewing the various requirements, a handful of potential locations were identified. What ultimately led to the site selection was the strong response that the client received from the local utility cooperative. The cooperative made it known that they wanted the data center on their grid and made commitments to keep key replacement parts in stock to quickly remedy if a transformer were to go down. They also negotiated a favorable utility rate. Having a cooperative, city or state willing to support a data center with financial and other incentives can be a huge factor in site selection.
Accessibility and Proximity
While it might seem to go without saying, accessibility is another key factor in connectivity. It is important that employees (and colocation clients) have access to “connect” with the facility. If a data center colocation facility is planned as the primary or secondary site, ensure that IT staff can easily access the facility. Everyone involved in maintaining the IT infrastructure must have easy access to the location. This translates to being close to metropolitan areas, transportation hubs (airport/train), major highways and other facilities.
Recently, we engaged in a discussion with a colocation client on this topic. Because of the nature of the company’s business, it was important that their facility not only be easily accessible, but also highly visible (to attract clients). Using their business goals and objectives as guides, we examined a number of facilities to determine what the risks were associated with each location to identify the location that offered the most “bang for their buck.” In this instance, the client was willing to assume the risks associated with putting the data center in a building in a downtown area next to a major interstate junction in order to increase accessibility to their clients. The company mitigated the location risks by creating an interior barrier before the data center entrance that in essence, created a building within the building. With no external wall, operations were protected from various situations such as a running into the building.
Proximity to a company’s other data centers is also another factor to consider in the site selection. On one hand, it might be beneficial to want to have the data centers close to each other so that employees are able to easily go between facilities. On the other hand, having two data centers in close proximity could present risk in the event of a weather event that could impact both facilities. Ultimately, it is a discussion that needs to be made based on the amount of risk the organization is willing to take on for their data center.
In another instance, site selection was constrained by connectivity requirements and necessary proximity to an existing data center. Specifically, the new data center needed to be within 50 to 200 miles to allow for active-active applications and ease of travel between locations. Additionally, given its Midwest geographic location, it was important that the two data centers were not located along east/west weather routes. The client didn’t want a snowstorm or tornado knocking out one location and then rolling through to the other facility. To tackle this challenge, we started by drawing a 200-mile radius around the existing data center. Then we researched weather patterns to identify the most likely progression of storms through the region. With that data in hand, we narrowed down locations by evaluating power and fiber optics within the remaining areas. Through this process, we lowered the potential risks while still addressing the client’s needs.
Evaluating Risk
As mentioned earlier, risk comes in many different forms with the majority falling into the categories of weather events, natural disasters and environmental hazards.
Weather Events. According to the National Centers for Environmental Information, “in 2014, there were eight weather and climate disaster events with losses exceeding $1 billion each across the United States.” From snowstorms to tornados, floods and hurricanes, weather is the number one cause of data center disruption and as a result, plays a huge role in site selection. While it is impossible to avoid weather all together, risk can be reduced by analyzing historical weather patterns as well as average rainfall amounts, average high/low temperatures and average wind speed/gusts. But if data centers are in “tornado alley” or an area pounded by snow each year, build redundancies into your system or constructing the facility with higher EF ratings to minimize risk as much as possible.
Earthquakes. Being near a fault line, for instance, could be especially troubling. Although there are building technologies that can mitigate risk, companies will likely pay a higher price to offset the cost of protecting the data center from earthquakes. Minimizing the risk from these factors ensures a more geographically stable location.
Railroads. If located near railroad tracks, consider the impact the vibration might have on the facility as well as the impact of potential derailments or containment spills that could force an evacuation. Unfortunately, much of the country’s fiber optic lines are laid adjacent to the railroad tracks because of the inherited easements. Buildings and equipment can be constructed to eliminate vibrations at a cost.
Transportation. For some, being close to a highway can be seen as an advantage even though the risks of containment spills, vibration and vehicle accidents are also present. It is a consideration to discuss to determine what the company and what company is willing to withstand in this area.
Neighbors. The operations of neighbors can also present data center risks. In general, select a location that is quiet without a lot of traffic. It goes without saying to avoid scrap yards, chemical plants and manufacturing operations where environmental spills can take place, explosives are used or activities might prevent facility access. If locating to a business park, the most important neighbors are the adjacent ones. When I evaluate business parks locations, I look for those that are nearly full or are well-established to know who the neighbors will be. In my opinion, a fresh park carries an added level of risk since unknown neighbors are also an unknown commodity.
Brownfields. When looking at sites within a metropolitan community – where connectivity is high and the infrastructure exists – it is unlikely that it will be a greenfield option. Unfortunately, with brownfields, the risks aren’t always apparent; sometimes they appear after construction has begun. In these situations, step back and reassess what risks your organization is willing to accept and how they will be addressed. This is a situation a client found themselves in just a few years ago. The client came to us and said, “This is the site we are going to build on – we already own it and it appears to be a good location.” Situated in an established business park, near an interstate in a large metropolitan area in close proximity to a major connectivity hub existed. The risk associated with this project weren’t originally apparent; it wasn’t until ground was broken that it was discovered the site was the location where the bad soil from the neighboring highway was dumped in the 1970s. The risks that were uncovered by the excavator weren’t something that could have been anticipated but it did require a conversation surrounding how to mitigate the situation so that the foundation was stabilized.
As you explore considerations when selecting a location for a data center, it is important to always keep in mind that there are tradeoffs that have to be made to achieve the optimum balance. Ultimately, our mission is to identify the requirements and the risks, explain the risks and figure out how to mitigate the risk to achieve maximum connectivity.
About the Author: Tim Kittila is Parallel Technologies’ Director of Data Center Strategy. In this role, Kittila oversees the company’s data center consulting and services to help companies with their data center, whether it is a privately-owned data center, colocation facility or a combination of the two. Earlier in his career at Parallel Technologies Kittila served as Director of Data Center Infrastructure Strategy and was responsible for data center design/build solutions and led the mechanical and electrical data center practice, including engineering assessments, design-build, construction project management and environmental monitoring. Before joining Parallel Technologies in 2010, he was vice president at Hypertect, a data center infrastructure company. Kittila earned his bachelor of science in mechanical engineering from Virginia Tech and holds a master’s degree in business from the University of Delaware’s Lerner School of Business.
