There have been many high-profile outages in public clouds, including those provided by, perhaps, the four most famous names in cloud computing – Amazon, Google, Microsoft and Salesforce.com. Table 3.1 below, which I compiled from numerous news reports, records at least 23 cases in two years from these four companies (Williams, 2010). Note that in some cases the outages were partial and did not affect all users, but if your company was affected this would be no consolation to you.
If outages would severely affect your business and you are confident your internal IT infrastructure is more reliable than
public cloud services then it may be advisable to only move non-critical business applications into the cloud. But are your internal IT infrastructures really more reliable than public clouds? According to IDC, the average mid-size company experiences 16 to 20 business hours of network, system or application downtime each year, which equates to 99.8 per cent availability (Boggs et al, 2009), while service providers such as Amazon, Google, Microsoft and Salesforce.com aim to provide at least 99.9 per cent availability and deliver on that promise for the most part (see again Table 3.1). Even with all the problems Google experienced with its Gmail service in 2008, which mainly affected a ‘small
number of users’, it was still available at least 99.2 per cent of the time for all users, and they guarantee at least 99.9 per
cent availability for Google Apps Premier Edition.