In the December 2009 survey we received responses from 233,848,493 sites, an increase of 212k since last month.

nginx continues to experience substantial absolute growth, increasing by 1.3M hostnames, which was the biggest gain this month.

Over the last 12 months, nginx has shown the most consistent growth out of all of the top Server Vendors, seeing only a small loss in February. Since this time last year, it has grown by 12.9M hostnames, which is only slightly below that of Apache's growth of 13.3M. In terms of market share, however, Apache has actually lost 4.7% this year, whereas nginx has seen a large increase, gaining 5.2% over the year.

In total we have seen an increase of 47M hostnames and 7M active sites over the last 12 months.

In November, Swishmail and Verio had the most reliable hosting company sites.

Swishmail is a New York based company that specialises in business email hosting. The company also offers several web hosting plans, all of which include access to a webmail client and shared calendars.

Verio is an NTT Communications company, using a distributed network of OEM and viaVerio channel partners to provide web hosting and SaaS applications. In September, Verio was named as the winner of the Best of SaaS Showplace (BoSS) Awards program.

Six of the most reliable hosting company sites in November were running Linux, including Verio, while both Swishmail and New York Internet were using FreeBSD. Netcetera provided the only Windows presence in November, with its site being served by Microsoft IIS 6.0 on Windows Server 2003.

Netcraft measures and makes available the response times of fifty 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.

Further information on the measurement process and current measurements are available.