As I stated in an entry last year, cloud storage will move out of its hype phase and into the acceptance phase when users don't even realize they're using cloud storage. For this to happen, ISVs are going to need to write applications that leverage cloud storage with the users having to interact directly with it. There are not many users who wake up one day and decide to go shopping for a cloud storage solution. Most have a problem to solve, one that can possibly be solved better by cloud-enabled software.
ISVs are leveraging cloud storage in numerous ways. Almost every backup application has been cloud-extended so that disaster recovery copies of backups can be placed off-site. Archiving applications, another obvious place to leverage cloud storage, have been the technology to store data that needs to be retained. There are several companies that use cloud storage as a globally available storage pool, allowing better collaboration among multiple users. Backup, archiving and file collaboration are early, cloud storage use cases and are growing in popularity. In most cases, the users are not interacting with cloud storage directly, but with an application, either one specifically written for cloud storage or one that has extended itself into the cloud.
The latest wave of cloud storage-enabled applications are those that leverage for more primary storage type of services, most notably SAN and NAS. While thought of as too risky just a few years ago, this market is now rapidly growing and producing revenue. In both environments, local storage acts essentially as a cache with cloud storage being the primary storage area. As data is updated it is replicated to the cloud, in the same way snapshot replication works on traditional systems. As data ages only the cloud copy is kept. A "cache miss" points the application or user to the cloud storage area, where data is copied back to the cache as it is being read by the user.
Using cloud storage in this way brings several benefits. First, worries about the cloud can be compensated for by having a large cache. Considering that percentage of active, in-use data at any given time is usually a fraction of the total overall storage, even if you have a cache twice as large as you active data set, physical storage costs can be reduced significantly. Second, data is automatically backed up to the cloud via the snapshot and then, usually, replicated within the provider's infrastructure. On the ISV products we have tested, the connection to the cloud and the cloud provider have all been up 100 percent of the time.