Cloud computing is the latest trend in hosted server infrastructure. As greater numbers of businesses outsource all computing to the cloud, network issues play a vital role in the successes of these initiatives. The use of cloud computing is cost effective and efficient as harried network administrators do more with less. By outsourcing server hosting and computing to the cloud, they greatly reduce the demands on their staff. It also makes it easier to back-up data and maintain updated versions. Just like many things about technology, there is no one size fits all definition of what is cloud computing. While some define it as any network activity done outside the firewall while others define it more narrowly around actual services offered. Regardless of the definition, cloud computing is a means of adding capacity and scaling up relatively quickly without having to invest heavily in infrastructure. It also removes the need for additional staff to handle the expansion and associated training costs. Whatever the reason for using cloud computing, having redundancy built into the infrastructure is essential. Cloud redundancy can be viewed from two angles. Firstly, the WAN connecting the user of the service needs to be redundantly connected. Without this, cloud users are vulnerable to not being able to connect to the cloud. WAN redundancy goes hand in hand with cloud redundancy and is an essential first step. Your hosted data is only as accessible as the reliability of your WAN connectivity. The second and equally important part of the cloud redundancy strategy is failover of the cloud itself. What happens if the cloud fails – the reason is not important – failure is failure. How do the users access their files, applications etc that are stored in the cloud. Redundancy has to be built in. Most cloud computing service providers take the precaution of copying customer data more than once. Each copy is stored in isolation and if failure occurs, only that part will fail and whole will not come crashing down. While this cloud redundancy is desirable, if the whole site goes down for example due to an earthquake, it is imperative that there is a disaster recovery site. This site – geographically separated from the primary site is the Disaster Recovery site. Stored data has to be replicated here continually so that customers do not lose anything should the primary site goes down. WAN redundancy is in the hands of the customer who controls the links to the Internet. Redundancy can be achieved by technology that exists in the market today. These devices accept multiple links while dynamically balancing the load. This single ended solution provides the users at the cloud connected location alternate paths to connect to the hosted services and applications. An end to end solution with dual ended redundancy solutions would have the ability to seamlessly handle failovers, failbacks and load balancing. Cloud redundancy is very important. Potential users should question providers what their cloud redundancy plan is and also the disaster recovery plan.