Cloud Downtime strategy: Building redundancy

Cloud services are not perfect. Even leaders like Amazon Web services have experienced a variety of downtime events. The good news is that using cloud computing, you can also somehow protect your services from the impact of cloud service provider downtime, which may not apply to traditional server hosting. Remember that the most important thing in your cloud computing downtime strategy is redundancy.

In addition to traditional backups, the most advantageous defense against cloud downtime is planning redundancy. The idea is simple: If a server goes down and other servers take over, the end user won't even notice the problem.

There are many techniques for implementing redundancy. One option is to put your server in a data center. Amazon Web Services (AWS), for example, lets you choose where you can host the server. In this case, a server can be placed in the data center of Virginia, where the redundant servers are placed in the Oregon data center. If your service provider does not offer a wide range of data centers, you should distribute redundancy among other vendors, and some servers run on AWS and some on Rackspace.

Next, prepare a management system and an appropriate infrastructure so that redundancy can really work in a downtime event.

Vcloud Director like VMware, Microsoft System Center, and cloud lifecycle management of the BMC can continuously monitor the server. If a server is down, you can remove it from the Active server pool until you find a backup. To do this, you need to configure your domain Name System (DNS) server so that when a client, such as a browser, seeks IP for a URL, its address can be returned to the addressed server. This enables the client to be active with the server;

Large enterprises are probably already using DNS configuration, but smaller companies do not use complex management software, but there is still a way to make it easier. Let the second server be ready to operate, but turn it off. Some cloud vendors do not charge for shutting down servers. In the event of a downtime, start a redundant server, go directly to your DNS manager, and convert the IP address to the new server. The effect may not be immediate, but with the current DNS server, you should be able to work quickly. The DNS Manager can issue an alert 24 small before a change occurs, but in practice, it is usually 15 minutes. To point out: if you maintain a traditional server power outage, you need to ensure that the latest version of the software regularly upgraded.

Take the right measures and plans in advance to help companies better cope with cloud downtime. If downtime occurs, end users will only experience a short downtime or not at all.

(Responsible editor: Schpeppen)