The Register has today reported that the ASA are pursuing Microsoft on their claim of 99.9% reliability:
The Advertising Standards Agency (ASA) is checking out a complaint about claims from Microsoft that it can guarantee 99.9 per cent uptime on its cloud services.
The Business Productivity Online Suite (BPOS) has been prone to outages. And even its successor, Office 365, has gone down twice since its launch in late June, leading some customers to dub it “Office 364”.
While Microsoft not the first company to claim ridiculous reliability, it’s frustrating to see another company guaranteeing the uptime that they cannot manage.
Of course, your website needs the best uptime and the best reliability for your customers, however; there are some important factors to consider
- 100% uptime is impossible to guarantee, there are too many factors
- scheduled downtime can be essential to ensure the security and speed of your site
- most companies who claim 99.9% or more uptime have probably never achieved it; and instead will pay compensation as a proportion of hosting fees by time
Often, if your site is unavailable under these ‘guarantees’; you must have noticed and claim for the compensation. Sometimes even requiring proof. Usually the amount paid for a few hours downtime in hosting fees will not compare to the amount of sales and confidence you may lose from your customers; especially at peak times.
We’ve worked with companies before who have claimed ‘Five Nines’ availability, with one notably using it as part of their identity, and ironically suffered with almost predictable downtime every week. So what is Five Nines? It’s commonly taken to mean “99.999%”, i.e. when the downtime is less than 5.26 minutes per year; here’s a short chart; shamelessly taken from Wikipedia’s entry on High Availability:
| Availability % | Downtime per year | Downtime per 30 days | Downtime per week |
|---|---|---|---|
| 99% (“two nines”) | 3.65 days | 7.20 hours | 1.68 hours |
| 99.5% | 1.83 days | 3.60 hours | 50.4 minutes |
| 99.8% | 17.52 hours | 86.23 minutes | 20.16 minutes |
| 99.9% (“three nines”) | 8.76 hours | 43.2 minutes | 10.1 minutes |
| 99.95% | 4.38 hours | 21.56 minutes | 5.04 minutes |
| 99.99% (“four nines”) | 52.56 minutes | 4.32 minutes | 1.01 minutes |
| 99.999% (“five nines”) | 5.26 minutes | 25.9 seconds | 6.05 seconds |
| 99.9999% (“six nines”) | 31.5 seconds | 2.59 seconds | 0.605 seconds |
However, like Microsoft, many of these sites can’t achieve anything like this uptime – instead often ignoring their ‘goal’, or occasionally opting to give compensation. However, a few hours downtime is usually only a few quid in hosting fees; when in reality it could mean hundreds or thousands in lost sales and confidence.
It’s clear that even with Cloud Computing, some downtime does happen. Both Google’s Docs and Apple’s iTunes have also suffered with unexpected downtime; and despite Facebook’s geolocated servers – they’ve struggled at times too. Even the Financial industry, arguably the most experienced in the business of reliability, has suffered with downtime previously.
Naturally new techniques have helped massively with reliability; but the general (mis)use of the term gives a very poor indicator. If you see a company advertising or guaranteeing uptime, remember – the compensation for downtime from a poor service does not compare to the sales that would’ve been gained using reliable hosting.
While it is possible to achieve an incredibly high amount of reliability – to be successful requires expertise, not compensation and outdated software.
P.S. Our uptime for the previous 12 months has been ~99.972%; with almost all of our downtime for scheduled maintenance and upgrade.