The cloud computing industry prides itself on the elasticity, flexibility and perhaps most importantly, high availability of its offerings. And yet cloud computing outages will happen, and when AWS, Google or Microsoft (News - Alert) experience downtime, it makes global news headlines.

TMCnet even discussed the top 10 computing outages of 2013, with the main offenders including Microsoft Windows Azure’s October 30 outage, which lasted more than 20 hours and slowed down global Internet traffic by nearly 40 percent. AWS experienced an outage on September 13 that lasted all of three hours and yet caused major disruptions and havoc. And NASDAQ’s noteworthy August 22 outage, caused by a software bug, disrupted U.S. trading for nearly three hours, a monstrous interruption for the financial community.

While cloud computing outages may be inevitable, there are things customers can do to plan and prepare for them to minimize downtime and manage their own customers’ experiences. Dell (News - Alert), which has been working diligently to beef up its portfolio of cloud computing solutions and services, recently recommended some tips to prepare for and combat outages.

According to Dell, a good cloud outage plan must take into consideration data recovery, continuity and, if necessary, infrastructure parameters. An outage plan will obviously be based on the user’s size and data consumption patterns as well as other variables. But most importantly, customers should know their cloud computing network in depth. This means that lists of dependencies, data access permissions and controls should be maintained and updated as part of an outage contingency plan. A map that outlines the core components for service recovery should be maintained as well, particularly if the customer is heavily dependent on the cloud for core services.

A good cloud outage plan will also address user experience issues and attempt to minimize disruption. Ideal strategies will give users access to alternative resources during an outage based on the parameters of their SLAs. Customers should also make an effort to control WAN traffic in the cloud in order to optimize network performance and direct users accordingly should a network-related outage occur. Data may well be dispersed among multiple locations with varying access times, which can slow network performance considerably. By having intimate knowledge of the WAN, customers may create efficient network paths among data sites to minimize disruption.

Finally, Dell recommends that cloud customers test their environments extensively, under real-world working conditions. This will ensure dependability and also create a paper trail to compare against when disruptions or outages occur.