What happens when a data center outage hits?
Recently, a very popular national newspaper experienced a data center outage lasting 15 minutes. During that time, viewers were unable to reach the company’s news website. I bet this wasn’t the 15-minute spotlight they were hoping to achieve.
As a data center guy, 15 minutes of downtime scares me because that can be a lifetime for many businesses. For e-commerce, media distribution, and transaction-rich organizations that rely on their IT infrastructure, 15 minutes has the potential to cost millions of dollars.
Why are uptime guarantees not enough?
However, N+1, 100% SLA uptime guarantees, and resilient IT infrastructure are often referred to as “checkbox items” when evaluating colocation providers. While generators and UPS systems may appear to be standard features of all data center providers, the end user must proceed with caution—because all resiliency is not created equal.
What should you ask before choosing a colocation provider?
Rather than assuming a provider has designed a facility and developed a service allowing for resiliency, you have a responsibility to your organization to ask the tough questions. Below are a few that are part of my colocation evaluation process.
1. What redundancy exists in the mechanical and electrical infrastructure?
Ask about the generators, UPS, transformers, PDUs, and chillers specifically. Anything less than N+1 at every level represents a risk to your business uptime.
2. Has the provider experienced any outages in the last three years?
What occurred, what was the impact, and why did it happen? Find out what preventative measures have been taken. The provider should be able to provide detailed information around each and every outage—whether it was service impacting or not—and offer a detailed explanation of how they plan on preventing this in the future.
3. How often is maintenance performed—and by whom?
Ask how often maintenance is performed on key components within their infrastructure, and who performs it. Maintenance should be performed on a quarterly (if not monthly) basis and executed by certified technicians.
4. What does the SLA really guarantee?
Ask about the SLA—what guarantees do they provide against downtime? My advice is to ignore the money penalties and instead look for a clear alignment between the provider’s interests and your own business continuity.
Why do the small details matter?
More often than not, we overlook the checkbox items in our evaluations and focus on the differentiators. Don’t get caught making that mistake—very minor differences can spell 15 minutes of disaster for your business.
