How Enseva Makes Colocation Better! – Part II

By Chris SeveyBlog

Last issue in Part 1 of this mini-series, I referenced some of the electrical and cooling enhancements that went into Enseva Next-Generation datacenter facilities and how these methodologies and technologies permit organizations to consolidate IT infrastructure into a reduced real-estate footprint. This provides an ideal environment for not only immediate cost savings, but long term cost savings as well. New high-performance compute infrastructure and continue to be swapped out during hardware life cycles without the need for increasing real-estate demands. I’d like to take the opportunity to re-evaluate some of these same technologies, but in a new light.

From time to time, we’ll be faced with disaster recovery type situations that require us to practically move mountains to help an organization return to operation. One significant occasion that comes to mind was the 2008 floods that devastated most of downtown Cedar Rapids. As the waters rose, organizations and people had to evacuate not only their offices and homes, but for some, their datacenters as well. I saw people on higher levels hanging portable generators out windows to keep a few essentially systems running. Others leveraged skyways to move equipment out of the pending flood zone. I was tasked with building out a new wing of a datacenter to help get mission critical systems back online for those fleeing the impacted areas looking for a new home for their servers, storage and network infrastructure.

Ultimately, a new 300kw Uninterruptable Power Supply (UPS) system needed to be purchased, installed and commissioned. Support infrastructure such as main electrical panels, branch circuit panels, conduit, wiring, cabinets, cable tray and network infrastructure all needed to be installed as well. The catch was that this all needed to be sourced, installed and ready for use in a single 48-hour weekend! I’m the type of person that thrives under high-pressure situations such as this, so this was my equivalent to an adrenalin pumping skydiving trip. Come Monday morning, I was a few hours shy on my beauty sleep, but every customer in needed of a new datacenter home had one.

Hindsight being 20/20 as it is, one thing learned from the floods was that expediting infrastructure and personnel to perform an impromptu datacenter expansion came at a hefty price tag. Overnighting large UPS systems, teams of electricians working around the clock and constant customer coordination for boarding all took its toll.

When setting out to architect the Next-Generation datacenter facility, one of the items needing to be tackled was a new method of deploying datacenter infrastructure quickly and in a cost effective manner. We wanted it to be as “cloud”-like as possible, where the premise of cloud is instant satisfaction by the delivery of resources just-in-time and as needed. Organizations should be free to define a datacenter environment that includes the quantity of cabinets, the type of electrical circuits those cabinets need and some method of network connecting the infrastructure to be installed in those cabinets with the rest of the organization. Then just push a button. Enseva’s Rapid Infrastructure Deployment (RID) allows us to accomplish just that.

In my field, electrical infrastructure in the form of branch circuits tend to be the most regularly occurring datacenter upgrade, so I’ll be focusing primarily on that during the blog post but will likely touch other aspects later down the road in other blogs. By definition, branch circuits are the electrical infrastructure from the breaker panel to the individual cabinets that you then connect cabinet Power Distribution Units (PDU’s) or sometimes referred to as power strips. Each time you need to deploy a cabinet, an individual with an electrical license needs to get power delivered from the live electrical panel over to that new cabinet without service disruption. These installs need to be scheduled, typically during a maintenance window, and after the necessary parts and materials are ordered and received by the contractor. Generally speaking, the typical turnaround time that I witness for non-rushed deployments are two to six weeks and largely dependent on the standardizations established throughout the datacenter and the relationship that has been established between the datacenter and an electrical contractor.

A few weeks ago an organization approached Enseva looking to establish a disaster recovery strategy that consisted of automated nightly backup replications to Enseva Hiawatha. This backup infrastructure would allow the organization to return to service using a standby environment should their primary environment fail. As part of the project, the resources required included cabinet space, electrical, cooling and a high speed internet connection to facilitate data replication. Thanks to Rapid Infrastructure Deployment, that customer was fully operational in under two hours’ time.

This deployment was made possible primarily through our non-arch flash breaker systems. When installing a new breaker into a live electrical panel, an arc flash (spark) is not created. This, among other technologies, is one of the factors that permit Enseva to rapidly deploy electrical infrastructure throughout the datacenter just-in-time. Gone are the days of scheduling and working around datacenter infrastructure timelines. Paying top dollar to teams of electrical contractors to deploy new infrastructure throughout the middle of the night are now a thing of the past. Most importantly, the next time disaster decides to strikes, everyone will be home in time for bed.

Some datacenters have gone down the path of busway systems in an attempt to achieve similar results, however in my opinion, there are some tradeoffs with such systems. First, most busway systems I’ve been in contact with lack security measures. For instances, there is no mechanism stopping a person from accidently or intentionally turning off the breaker switch that’s exposed to the world. This is something that simply couldn’t be tolerated in a multi-tenant facility. This concern was rectified in Hiawatha by placing breakers in a secure housing ensuring breakers are only accessible by datacenter staff. Second, busway systems often need to be extended multiple times to be useful. Each time the bus bars are extended, you have joints connecting the individual units together. These joints equate to points of electrical failure and downtime. Enseva Hiawatha instead chose to leverage short, seamless buss bars and extend electrical whips to customer cabinets, yielding a more reliable method of electrical distribution with fewer points of failure.

Stay tuned in next week when I plan to discuss the additional benefits that Enseva’s electrical system provides to its users by brining monitoring, trending and pay-as-you-grow to the masses. In the meantime, if you’d like to learn more about how Enseva’s making a difference in the datacenter world or check out our rapid infrastructure system, schedule a time to stop in for a personalized tour by emailing us: