How resilient is your data centre provider?

Since we’ve mentioned tier classification, let’s look first at one of the criteria this is judged on: the presence of redundant infrastructure for power. Within Tier 3 (claimed by the majority of commercial colocation facilities) UPS redundancy can be applied to facilities with N+1 right through to 2(N+N) resilience. That covers a wide range of possible configurations – with associated variations in the degree of protection that’s offered in terms of keeping the power flowing to your kit. Find out from your provider where they fit within their claimed Tier.

Where does the power supply to the facility come from? Is it from a spur (due to cost or the availability of land in a built-up area) or a ring (greater redundancy, since mains power can be delivered from two independent substations – critical infrastructure sites are built on rings for this reason).

It’s also a fact that mains failure happens. When a power cut occurs, how much time is available on UPS batteries and on-site generator fuel supplies to keep your business going?

If there’s lots of power, feeding lots of equipment, there’s going to be lots of heat to deal with. If infrastructure isn’t cooled within a data centre environment, it won’t last long, which is why Tier 3 facilities are required to maintain high capacity air conditioning units with a minimum of N+1 redundancy. You would also expect to see hot or cold aisle containment systems to help with efficient environmental control.

If power and cooling are suitably catered for, what about the other critical aspect of keeping your business up and running: a connection out to the wider world? Consider what would happen if your network provider suffered an outage due to a physical fault such as a fibre cut? Does the facility offer scope for adding resilience to this aspect of your operation? What opportunities are there for diverse transit, and how would this work in practice?

Judgement shouldn’t be restricted to power, cooling and network either – looking beyond infrastructure resilience is an important consideration. How long has your provider been around? What do their investment plans and financials tell you? Are they likely to be around in five year’s time? Anyone who knows how long a data centre migration takes will understand the critical importance of choosing a provider with long term stability. Try to be as sure as you can be that any future migration will be your controlled decision, and not a forced, emergency relocation in the event of your provider’s demise.

A data centre with high levels of infrastructure redundancy is still vulnerable to downtime if it is not supported by a knowledgeable, process-conscious team. And no matter how many banks of UPS batteries are available, business continuity and uptime will be at risk if equally high standards of security at the facility are not upheld.