What to watch out for when choosing your hosting service provider.
As any new technology becomes more commonplace, to the point of commoditisation, price starts to become one of the main points of differentiation. You’ve only got to look at the way mobile phones are sold to see that in action on a daily basis.
It’s also present where cloud hosting is concerned. No one would argue that price isn’t important; it is, rightly, a vital commercial consideration of any purchase. But when price is the focal point of the conversation it can be all too easy to lose sight of the combination of factors that go into determining that price. Particularly when it comes to technology that, because of its utility-like nature, is almost taken for granted.
It’s all too easy to switch off a little and become ambivalent to the things we are familiar with, and so it never hurts to be reminded not just of what something does, but how the nuances of its performance, and pricing, can affect the rest of our working lives.
Tiers are more than just stepping stones
That’s certainly the case when it comes to data center tiers. This standard methodology for defining data center engineering standards, in tiers one-to-four, lends itself too easily to being dismissed as superficially obvious. After all, Tier 1 is the least well performing and Tier 4 is the best, Tier 1 the cheapest option and Tier 4 the most expensive right? Once you’ve got your head around that, what else is there?
Therein lies part of the problem with taxonomies; easy to read labels can hide a wealth of valuable information.
Getting your head around some of the differences in the tiers will mean you’re able to make better decisions when it comes to choosing a data center, and most importantly you’ll avoid a number of potential pitfalls.
Crunching the numbers
First let’s consider the uptime differences between the four tiers.
Tier 1 equates to the basic requirements for a data center – a single distribution path for power and cooling serving the processing hardware, and non-redundant capacity. Annually, uptime would be at the rate of 99.671% and include 1,730.4 minutes (1.20167 days) of downtime.
Tier 2 infrastructure gives you redundant capacity that can be swapped in and out without causing system failure. It will also include one non-redundant distribution path. Tier 2 data centers are not intended to be fault tolerant when it comes to things like natural disasters, but they have 99.741% uptime, and 1,362.2 minutes of downtime, per year. That equates to almost 23 hours.
Tier 3 data centers have one active and one passive distribution path. They’re built to be more resilient, although they’re not intended to be completely immune to disasters. But they will withstand them better, thanks to higher capacity equipment. Tier 3 data centers offer 99.982% of uptime and 94.7 minutes of downtime per year.
Tier 4 are fault tolerant with two active distribution paths, offering the highest level of capacity, including mission-critical levels of tolerance, fully redundant subsystems and components with concurrent maintainability. They also have dual-powered cooling systems. Uptime is 99.995% per year, and the annual downtime is 26.3 minutes.
Get the balance right for your business
Unless you are running genuinely mission-critical applications and downtime is utterly unconscionable, you are unlikely to want to shoulder the not insignificant investment a Tier 4 data center will require, which could be as much as double the cost of a Tier 3. Tier 4 also involves a very high environmental overhead, and a much higher carbon footprint.
Similarly, unless all you’re doing is running a simple information-only, non-transactional site, you should walk away from Tier 1 too.
Identifying which of the two remaining tiers is right for you comes down to understanding the impact of some of the differences.
Even at the headline level, the difference in annual downtime between Tier 2 and Tier 3 is substantial, from almost one full day, to just over an hour and a half, respectively. Are there going to be cost implications? Of course there are. But pause, if you will, and reflect on the difference in those downtime numbers; one is the time it takes to fly from London to Sydney. The other is the length of a football match, with a little stoppage time added on.
That’s a considerable difference, and you’d need to think carefully about the potential for exposing your business to harmful consequences as a result of it. A lot can happen in 24 hours.
If you have SLAs in place with your clients and customers that include penalties and reparations for missed application availability, you need to map that against what's being offered by your data center. If unexpected outages prompt modest financial reparations from your data center, for example, how does that stack up with the potential losses for the business caused by those outages? Will it matter if customers come to your site and can’t transact with you? Will your clients invoke penalty clauses in SLAs if applications they depend upon are impacted?
Making the right choice
Weighing up which data center tier will be the best fit for your business operations is more than a simple price vs uptime equation. Price differentials are inexorably lined to tiers, but you can’t do a like-for-like comparison in these scenarios unless you are fully up-to-speed with all the implications. You need to peel back some layers and understand what the implications are in the event the lights go off.
Hope for the best, prepare for the worst? Maybe. But be aware of the need to do some risk assessment when it comes to deciding where your business will be hosted. Knowing which data center tiers your potential providers operate from will always be an essential first step.