|2||Qube Managed Services||Linux||0:00:00||0.000||0.138||0.058||0.116||0.116|
|9||Hyve Managed Hosting||Linux||0:00:00||0.009||0.091||0.059||0.122||0.122|
Swishmail had the most reliable hosting company website in August 2017. It responded to all of Netcraft's requests from several points around the internet, with not a single failure, and an average connect time of 57.6 milliseconds. The US-based company provides email and web hosting services based on FreeBSD, and this is its fourth appearance in the top ten so far this year.
There was an incredibly close battle for first place this month, as Qube Managed Services also managed to respond to all of Netcraft's requests, but its average connect time was fractionally longer at 57.7 milliseconds – less than 0.2 milliseconds slower than Swishmail's average. This puts Qube in second place, but only just. Impressively, this is Qube's seventh time in the top ten for 2017. It operates data centres in London, Zurich and New York, where it offers managed hosting, colocation, and cloud-based solutions.
In third place, with a single failed request, was One.com. Its average connect time was noticeably faster than those of Swishmail and Qube; but reliability trumps speed in this contest. This is One.com's fifth top-ten appearance this year, including another first place ranking in April. Founded in 2002, the company was established in Denmark, and offers a variety of internet services including 1-click WordPress installations, domain registrations, email hosting, and Office 365.
Five of this month's top-ten hosting companies used the Linux operating system for their main sites during August, while two used FreeBSD, one used SmartOS, and another used an F5 device. Conspicuous by its absence is Windows, which has not appeared in the top ten since January, despite being found on more than half of the 1.8 billion websites in our Web Server Survey.
Netcraft measures and makes available the response times of around thirty leading hosting providers' sites. The performance measurements are made at fifteen minute intervals from separate points around the internet, and averages are calculated over the immediately preceding 24 hour period.
From a customer's point of view, the percentage of failed requests is more pertinent than outages on hosting companies' own sites, as this gives a pointer to reliability of routing, and this is why we choose to rank our table by fewest failed requests, rather than shortest periods of outage. In the event the number of failed requests are equal then sites are ranked by average connection times.
Information on the measurement process and current measurements is available.