There is an arcane but important practice for protecting data called Backup. Backup began in the early 60’s with the creation of redundant digital copies of data. Today, it is practiced primarily as a discrete utility based on ‘backup software’ making copies of data to tape, a reliable and low-cost removable media.
Tape backup is a pervasive technology. Currently, more data resides on tape than on disk. However, tape backup systems are surprisingly unreliable. In large datacentres tape backup is the source of hundreds of errors every day. In an attempt to improve backup integrity, businesses have implemented a variety of complicated tape backup systems, including large-scale tape automation.
These large-scale systems require specialist administration by system experts, skilled in the art of operating tape backup. Even so, with first time tape backup success rates in the region of 70% to 80%, data protection is constantly compromised, and as the backup window closes, in the current regulatory climate, this is a cause for concern for any business. Given the demonstrable unreliability of tape and the combined infrastructure and human resource costs needed to manage the systems, tape as backup starts to become a dubious value proposition in the datacentre.
Businesses would never accept an operating environment as poor as tape for primary storage and are struggling to justify it even for backup. Organisations are beginning to look at solving the problems associated with tape backup not by installing better automation or simply throwing more money at it. A fundamental rethink of how data is stored has begun.
The very need to backup data as a separate exercise is being called into question. No matter what approach is taken, backing up data will always be complex. Anything that is complex almost inevitably fails. IT has to ensure data integrity, reduce operating costs and cope with a smaller staff. To solve the complexity issues surrounding the back up of data, some organisations have, simply, decided to stop backing up.
Moving away from backup methodologies requires a fundamental shift in redundancy architecture, moving to a simple, transparent operation where redundancy is native to the write process, (i.e. some form of replication with versioning) where data is always there, and even the concept of “restore” goes away. The failings of tape backup are driving the revolution. Cheap disk-based storage is an enabler and accelerator of the transition.
With the simplification of the storage management problem described above, the Data Management Forum’s (DMF) ILM Initiative has conclude there is a need for a new approach to data management. The DMF is putting forward a proposal to the SNIA to define a vision for ILM as, “A new set of management practices based on aligning the business value of information to the most appropriate and cost effective infrastructure.”
This proposal, if acted on, will go beyond the SNIA’s charter. The tactic the DMF is taking is to focus on the storage and data aspects of this model while working with other associations. The real power of this approach is that it will unify the management of network, computing, and storage infrastructure over time.
Storage by tape will remain essential in the overall picture, but will be relegated to archive and long-term retention.
Nexsan technologies is exhibiting at Storage Expo 2006