See full table

Rackspace had the most reliable hosting company site in October 2021, with an average connection time of 8ms and no failed requests. So far in 2021, Rackspace has appeared in the top 10 every month, with four appearances in first place. Rackspace provides a wide variety of cloud services from its global network of over 50 locations across five continents.

In second place, CWCS Managed Hosting also responded to each of Netcraft’s requests in October, with an average connection time of 62ms. CWCS provides dedicated servers and cloud services with data centres across the UK and North America.

The top five is completed by Choopa.com, Bigstep and Swishmail whose sites each responded to the same number of requests and were separated by average connection time. Choopa.com had the fastest average connection time of just 3ms. Choopa.com offers a range of services including cloud hosting, dedicated hosting and colocation with its own primary facility in Piscataway, New Jersey and other facilities in Los Angeles, Amsterdam, and Tokyo. Colocation from Choopa.com is available in the US, Amsterdam and London.

Bigstep provides bare metal hosting solutions from data centres in the UK and Romania with additional facilities in North America and Europe for project delivery. Swishmail provides business email services with hosting services from three data centre facilities in New York.

Nine of the top 10 hosting company sites used Linux in October, continuing the dominance of Linux. In seventh place, New York Internet (NYI) used FreeBSD.

Netcraft measures and makes available the response times of around twenty 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.