Companies that choose to use cloud services have certain expectations. They expect to save a little bit of money, as they don’t have to install expensive on-premises equipment. They expect that service providers will have a knowledgeable, experienced IT staff that can address any potential problems and keep the services running without glitches.
And speaking of keeping services running, companies that take advantage of cloud offerings expect that the services will be available all day, every day, without significant outages.
In the past two weeks, companies have had to deal with major outages by Salesforce.com (News - Alert), one in late June and one on Tuesday of this week. This most recent outage affected nearly all of the 68,000 Salesforce.com customers for at least an hour. The 10-year-old Software-as-a-Service provider, the largest on the market, suffered a systemic data center failure that disrupted both normal operations and backup systems.
Salesforce’s cloud-based CRM systems are divided into many instances around the world, and each instance serves a distinct geographical region. No less than seven of these instances went down on Tuesday, beginning with NA1, NA5 and NA6 in North America. When NA6 went down, Salesforce.com’s Application Store also failed, as it shares infrastructure with that instance.
Soon CS0, CS1, CS3 and CS12 went down as well. These instances are part of a “sandbox” in which Salesforce.com users can develop, test and preview new features. The outages began just before 4:00 a.m. EDT and in many cases persisted for almost six hours. The cause of the outage was not immediately clear, though initial comments from Salesforce.com indicated power problems.
This problem comes on the heels of the June 29 outage which primarily affected the North American instances, starting in NA2. Salesforce.com blamed a fault in its storage tier, and the problem eventually spread to a European instance. All told, the outage lasted for approximately six hours.
Earlier, on June 14, Amazon Web Services also suffered a significant outage in its Northern Virginia data center, knocking out websites such as Quora, Hipchat and Dropbox (News - Alert). Heroku, a Platform-as-a-Service provider owned by Salesforce.com, was also affected. This outage was caused by a cable fault in the high voltage power distribution system and lasted for over 18 hours.
Ultimately, millions of people rely on these cloud-based services, and may companies count on them for vital business processes. The cloud is a viable option for most because it ensures reliability – you don’t have to have your own IT staff to deal with potential problems, you can rely on the provider to take care of issues for you. But there are many who are reluctant to jump into the cloud for the very reason that they don’t have full control, and instances like these outages are what give people pause. Of course, disasters and accidents will happen, but each minute of downtime costs money and creates headaches and hassles, so these providers need to take steps to ensure that outages like these do not happen in the future. After all, if they don’t, someone else will, and they will be left in the dark, not because of power problems, but because of defecting customers.
Want to learn more about cloud communications? Then be sure to attend the Cloud Communications Expo, collocated with ITEXPO West 2012 taking place Oct 2-5, in Austin, TX. The Cloud Communications Expo will address the growing need of businesses to integrate and leverage cloud based communications applications, process enhancement techniques, and network based communications interfaces and architectures. For more information on registering for the Cloud Communications Expo click here.
Stay in touch with everything happening at Cloud Communications Expo. Follow us on Twitter.