15th March 2017
One of the key considerations when choosing web hosting is deciding what it is worth to your organisation for your website to be available. How much of the time does you website need to be available?
The obvious response is that 100% availability is required. In fact, the situation is usually a little more subtle.
100% availability is always the starting point in terms of the uptime we expect. In the real world, there will always be small factors that might impact upon that. Over a long period of time, wear and tear of hardware, maintenance time and power cuts might have an impact on a single server in a single geographic location.
As such, here at Tibus we’re able to guarantee availability of between 99.982% and 99.995% for a server in our Tier 3 or Tier 4 classified data centres.
Is that good enough for your organisation?
The next question to ask is, can your organisation tolerate potential downtime of between 0.018% 0.005%? To put that into context, availability of 99.995% equates to a maximum downtime of 10-and-a-half hours over the course of a year.
Would your organisation be able to accept unavailability at that level? For a service business fielding customer enquiries mainly during normal office hours, the answer is probably yes. For a health organisation that might need to access data to make life-or-death decisions during that time, the answer is undoubtedly no.
A couple of extreme examples, but you will take the point. At this stage, it is just a case of balancing the cost of putting in measures to further improve levels of availability with the cost to the organisation of not having those measures in place.
100% availability
If your reputation depends upon being available at all times, if there are serious commercial implications for even a few minutes of downtimes or if you simply believe, as we do, that every customer ought to be able to visit your website whenever they choose to do so, you might opt for a hosting setup that pushes you closer to that 100% availability.
We usually achieve this through dual-server or multi-server architecture and load balancing. In other words, you have more than one server so that even if maintenance and updates (which often requires a server reboot) are required, another server can pick up the slack.
There can also be performance benefits, in terms of page load time, for your customers when you spread the use of your server’s resources over multiple servers.