Data Center vs. Colocation: Fly Solo or Go Colo? Know Your Options

Here's a primer on what to consider in determining the best solution for your clients' data needs.

July 18, 2019

6 Min Read
Go solo
Shutterstock

By Marc Cram

Cram-Marc_Server-Technology-150x150.jpg

Marc Cram

Resellers are often speaking with a team of top mechanical, electrical and system engineers and project managers who work on data center projects. From pre-deployment data center site surveys through deployment services and customizable support options, you’re challenged with knowing exactly what customers need.

On the flip side, your customers realize that successfully building a data center depends greatly on the suppliers and partners chosen to help in the design process. Positioning yourself as the most innovative company providing a broad line of data center specific products — with the best feature sets needed to support customer goals — will increase your chances of securing enterprise or colocation business.

Helping your customers select the best solutions to support their data center choice, whether they build a data center (fly solo) or are heading to a colocation data center facility (go colo), doesn’t have to be difficult — if you know the correct lingo to use.

Flying Solo

If your prospects are going solo, they’re building a data center and need a solid team of professionals working together. Knowing IT systems and software isn’t enough to convince a customer you’re the most qualified channel partner to consult on building a facility. Channel partners should be prepared to intelligently speak about finance, utilities, project management and building management.

The complexities of today’s data centers present a formidable obstacle to many resellers. Below are a few of the numerous questions your customers are considering, ones your company can leverage for your product and service positioning:

  • What is the nature of the computational load in the data center? Does it require a particular amount of uptime (99.999%) or Tier Rating (as defined by The Uptime Institute)?

  • Is the data center expected to achieve a certain efficiency — its power usage effectiveness (PUE) — or carbon footprint?

  • How much power and cooling per square area will be required?

  • How is floor space to be allocated between actual compute infrastructure and all the other support around it (offices, cooling, etc.)?

  • Will the data center require 24-hour onsite personnel support?

  • What are the goals for latency/responsiveness from the data center?

  • Will the data center need to have a secondary disaster-recovery facility built to run in parallel?

  • How quickly does the business need to have systems up and running?

  • What is the supply base ecosystem that you will need? Open Compute vs. Open19 vs. OEM vs. ODM?

  • How long is the data center expected to be in operation before a refresh cycle occurs?

Going Colo

Colocation data centers rent rack space to third parties for their servers or other equipment.

In theory, redeploying existing IT assets should be a low-risk path to success. Your customers know things already work together and everything is already in a rack, so how hard could it be? In addition, your customers are thinking that deploying new assets to the colo means they can start fresh. Here is the opportunity to position new products and new architectures that are likely to be more efficient and more reliable than the aged gear they already own.

But for some companies, choosing to go with new hardware is a headache. This presents an opportunity for resellers because to minimize that pain, these companies will be looking for …

… best-of-breed suppliers that can help pull together a range of products and services into a “guaranteed to work together” solution. Businesses will be looking for resellers with a broad offering to support the infrastructure, no matter where on the globe it will be located.

Before businesses make the jump into a colocation deployment, resellers should know and present these factors to their customers:

  • Can the colocation data center handle your rack power density today with a single power drop to the rack?

  • What about for future needs?

  • Can the colo provide adequate cooling for the targeted rack power density and temp specs, today and for the foreseeable future?

  • Will you rely on the “remote hands” IT management service of the colo provider or some third party to reboot locked-up hardware?

  • Is the IT load time-sensitive or highly impactful of revenue?

  • Will the racks in the colo require frequent access, whether for changing IT configurations or repair purposes? If yes, by whom? Would it make sense to have an asset management system in place at the rack level?

  • Does the IT hardware support the Intelligent Platform Management Interface (IPMI) or does it require other out-of-band (OOB) access?

  • How best to achieve multiple paths to the control interfaces of remote hardware?

Power Is Consideration No. 1

Whether IT gear is located on-premises (flying solo) or off-site in a colocation facility (going colo), power should be one of the first discussion items. Why? Because power is one of the highest recurring costs when operating a facility and is always on a customer’s mind. To alleviate this concern, choose a power distribution unit (PDU) that can help minimize the number of power drops. This will reassure your customer they’re receiving and making efficient use of all the power they pay for each month.

Power devices are critical and must have the right form factor, outlet count, and feature set for specific applications. It’s also important to consider whether customers need 1kW, or over 100kW in a PDU, and AC or DC power. To support this, selecting the best PDU is crucial to achieving operational uptime and efficiency goals — two very important power considerations for solos or colos.

Today’s workloads need to be supported with remote outlet power measurement and control and intelligent PDUs that offer the convenience of remote reboot via an embedded web interface or through SNMP as well as the ability to monitor:

  • Rack-level infeed power

  • Outlet power

  • Rack access via IP camera

  • Environment including temperature, humidity, air pressure differential and floor-level water detector

These monitoring abilities are in addition to helping with asset tracking, smart load shedding and capacity planning needs.

Positioning yourself as the best-fit reseller to support a company’s data center choice, whether to fly solo and build a data center or go colo and move IT assets into a colocation facility, doesn’t have to be difficult — if you know customers’ questions in advance. Like a chess game, being several moves ahead, anticipating pain points and intelligently alleviating them will ensure the win and repeat business.

Marc Cram is director of new market development for Server Technology, a brand of Legrand (@Legrand). A technology evangelist, he is driven by a passion to deliver a positive power experience for the data center owner/operator. He earned a bachelor’s degree in electrical engineering from Rice University and has more than 30 years of experience in the field of electronics. Follow him on LinkedIn or @ServerTechInc on Twitter.

Read more about:

MSPs
Free Newsletters for the Channel
Register for Your Free Newsletter Now

You May Also Like