After Google's email service temporarily disappeared this week, Check out the worst examples of web sites and internet services crashing - here's the top 10:
Millions of internet users worldwide sit up and take notice when Google goes down. The latest outage on Tuesday 24 February lasted for more than two hours, with the web giant attributing the unplanned downtime to testing new software during datacentre maintenance. Google has promised that the minority of subscribers who were paying to use Google services at the time will receive 15 days of free access. On 11 August 2008 Google’s Gmail site went down for a couple of hours, again inconveniencing millions of people.
1901 Census site
The Public Record Office made the mistake of advertising its 1901 census website heavily in advance of its launch, thereby stoking unprecedented levels of interest in genealogy and UK ancestry around the world. Not only did the site promptly crash under the weight of an estimated 30 million visitors a day (it was designed to handle a daily influx of 1 million visitors), it was withdrawn five days after its official opening in January 2002 and stayed down for a further seven months.
Snow trains today
February 2009 saw many travel web sites flooded with additional visitors as hundreds of thousands of stranded workers struggled to find out if they could get to work. The Transport For London web site was briefly down, while the National Rail Enquiries web site ran at "reduced pace" as hits rose 800 per cent compared with a normal Monday, with more than 32,000 users visiting it every second. South West Trains also reduced its homepage to a bare minimum and posted a note about heavy web traffic, while enquiries on the AA and RAC web sites were also well above average, slowing down access considerably.
Severed underwater cables
In December 2008, a severed fibre optic cable under the Mediterranean Sea broug ht severe disruption to countries in the Middle East for over two days, with up to 70 per cent of all internet traffic and telephone communications between Europe and Africa affected. Internet traffic had to be rerouted through Asia and the US to keep people connected. A similar outage halted communications between Europe, Africa and Asia earlier in 2008, believed to have been caused by ships’ anchors ripping through a section of the same cables.
Not so Hotmail
The last big global Hotmail outage, reported to have also affected MSN messenger and other Microsoft Live services, happened this time last year (February 2008) and lasted for over two hours, with another big outage in 2006. MSN Messenger was reported to be down for up to a week for some users in July 2001, with Microsoft blaming an extremely rare set of circumstances resulting from one of its database servers having a disk controller failure.
YouTube IP hijacking
Video sharing web site YouTube was put out of action for two hours in February 2008 by a Pakistani telecommunications company, which was attempting to block Pakistan residents from accessing the site. The telco was blamed for passing on the IP addresses to other Pakistani service providers so that YouTube traffic could be redirected elsewhere, where a network configuration error led to them being blocked by many of the world’s ISPs.
Skype reaches the limit
In August 2007, a software fault prevented up to 220 million registered users from ccessing and making internet telephone calls using every version of Skype software downloaded since 2003. Skype, acquired by Ebay for $2.6bn (£1.8bn) in 2005, said at the time that it did not fully understand the reasons for the outage, but denied that it had anything to do with capacity problems caused by its rapidly expanding user base.
Amazon S3 suffers
Amazon’s cloud computing platform Simple Storage Service (S3) is supposed to provide users with storage on demand, which it does – unless the site is down. It was hit by an outage in February 2008, as one of its datacentres became unreachable when the maximum capacity of its authentication service was breached, according to the company. Fresh outages were suffered in June of the same year amid speculation that Amazon was targeted by a denial of service (DoS) attack.
Parts of the world’s best-known online auction site was down intermittently for almost a day in June 1998, as a result of a failure in the software used to list items for sale and update bids for the outage. Ebay is estimated to have lost over $3m (£2.1m) in revenue due to customer refunds and waived fees, and had to extend auctions to make sure bidders were given a fair deal. Periodic problems in 1999, 2002 and 2003 followed, leading Ebay to formulate an outage policy for customers and seek to reassure investors that reliability issues had been resolved.
Going down in San Fancisco
In July 2007 a major electrical outage in San Francisco saw a 227,000 sq ft datacentre hosting popular sites such as Craigslist, TypePad, Technorati, Second Life and Gamespot go down for several hours, demonstrating the importance of effective disaster recovery strategies.