It's no secret that the old-fashioned approach to data protection – backing up copies to tape -- often can't keep pace with the disproportionate information growth faced by many companies.
Not surprisingly, tape-based backup solutions are losing popularity due to many shortcomings, including lack of flexibility and scalability that numerous emerging data protection solutions are trying to address.
Recently, customers started to examine a number of different approaches to data protection, some of which extend the tape-based backup model, such as VTLs (virtual tape libraries), while others propose a completely different approach.
Obviously, the advantage of using a VTL is to expedite backups and maintain the same tape paradigm without disruptive changes to operating procedures. In most cases, with tape virtualization a company can use the same backup application and tape technology. Virtual tapes grant faster and more flexible backup and restore cycles.
One of the latest VTL solutions to enter the North American market, CentricStor, was recently ported from Europe where it was developed by Fujitsu-Siemens. CentricStor targets companies with large requirements, and has numerous features to accommodate applications running on mainframes or open systems. Recently, the company added support for more mainframe OSes, including IBM OS400, VM, and VSE (Virtual Storage Extended).
I won't even try to give a comprehensive rendering of its features in this limited space, but CentricStor supports major backup applications and can connect as many as 32 real tape drives, including IBM, StorageTek, and LTO (linear tape open) models.
In between those hosts and the actual tape libraries, CentricStor offers redundant RAID arrays with hundreds of virtual tape drives managed via built-in applications.
Add system partitioning, cache management, compression, and a variety of reporting tools to the above features and you have a system that has little in common with an old-fashioned tape library.
Like other VTLs, CentricStor essentially replaces old tape media and drives with more flexible and faster virtual images. The amount of data to back up, however, remains the same.
Other vendors are taking a different approach that promises the same data protection of traditional backups while minimizing the amount of data copied during each backup operation. Let me explain how this works.
In most databases and users' files, the amount of new information is only a tiny fraction of the whole archive. For example, think of a customer roster or of an order database. Backup applications see only one side of the backup scenario, the source, ignoring the fact that most data has already been protected by previous cycles.
As a consequence, each backup operation copies an enormous amount of data that isn't really needed, because it's already present on the backup media.
Data Domain, a startup company, has developed an innovative backup solution that cuts dramatically the amount of data transferred at each operation by removing redundant data from each backup. The solution from Data Domain replaces tape drives with a smart disk array, called the DD200, an intelligent appliance that analyzes the backup stream and removes redundant, already-protected data. The DD200, which has been around since the end of 2003, works with most backup applications and offers an average reduction in the amount of data comparable to a 20:1 compression ratio, according to Data Domain.
Last week Data Domain began shipping Replicator, an application that extends the same frugal data transfer algorithm to remote connections. Replicator is an asynchronous mirroring application that keeps two DD200s automatically in synch over a WAN. A journal system shelters the application from WAN hiccups. According to Data Domain, Replicator can pump the equivalent of 50MBps of native data, although the mileage may vary depending on the data content and the connection speed.
Obviously, CentricStor and the DD200 plus Replicator combo address different customers and problem areas, but both preserve your favorite backup application while removing some of the most annoying limitations of that old backup paradigm.