The reinvention of backup and disaster recovery
As data volume continues to grow exponentially and additional requirements for retention and availability of data for compliance purposes have been heaped upon companies of all sizes, traditional backup, recovery and disaster recovery methods have become progressively more painful for IT managers. Backup data set volume, by its very nature, typically takes up 10 times or more the storage capacity of primary data, which tends to make it immobile. Much like the oversized presentation that, 10 years ago, you found you couldn’t e-mail to anyone, huge backup data sets in effect become albatrosses. You can’t move them efficiently via WAN, so they get offloaded to tape and relegated to physical handling and transportation by trucks. With an increase in hacking, natural disasters and concerns over terrorism, there are more threats to the enterprise than ever before. When you consider that the elapsed time during which the typical business can survive without its IT function is narrowing across all industry sectors, protecting data assets has never been more important. Yet the way most companies have had to go about data protection is with tape, which is difficult to effect fast and reliable restores from and carries inherent security risks (3rd party handling, loss, mechanical failures, human error during tape retrieval, etc.) when used for disaster recovery purposes. All this points to a very real need for a paradigm shift in the way companies approach their data protection. However, before entertaining some of the available new technologies and solutions for backup, recovery and remote disaster recovery, the problem that lies at the root of all issues needs to be acknowledged and resolved: massive backup data volume must be reduced significantly. The respective ‘reinventions’ of the backup/recovery and disaster recovery processes are all contingent upon reduction of backup data volume. Deduplication: the key enabling technology First, data that has been reduced to this degree can cost effectively be stored on disk instead of tape. In fact, in many cases it can actually be stored at a significantly lower cost per Gb than tape. Disk-based backup is inherently more reliable for restores than tape - and it’s much faster to restore. With deduplication and disk-based backup, backup windows (increasingly a pain point as primary data storage capacity requirements grow) are shrunk significantly. But deduplication’s role in solving data protection problems is elevated even further in disaster recover applications. Once data is reduced, data mobility is possible for secure and reliable remote replication via WAN. Beyond the data volume reduction already achieved, deduplication enabled DR topologies work to ensure that redundant data that already exists at the remote DR site does not get transferred, so network efficiency levels are increased by up to 99 percent. This means that bandwidth costs are kept in check. In fact, most users find their existing network infrastructure can now accommodate backup data transfer – with no expansion in bandwidth required. Finally, by keeping backup data on-line (both locally for more extended periods and remotely for disaster recovery), the management headaches and administration costs associated with tape and traditional ‘tapes on trucks’ scenarios is completely eliminated. The entire process can be centrally managed and automated on-line. In larger multi-site topologies, consolidation to tape may be desirable for archiving purposes, but tape’s role as an active part of the back-up/restore/DR process can be conclusively put to out to pasture. The principles behind data deduplication technology and its application benefits are simple enough to understand, and hundreds of companies have now availed themselves of the benefits and topologies enabled by deduplication. But viable deduplication technology is really hard to create. Performance, data reliability and system scaling considerations separate effective solutions from the pack of vendors that currently offer data protection systems. For those interested in pursuing deduplication enabled back-up and DR solutions, here are a few tips to get you started in your search: * Beware the compression ratio stat hype. Different types of data compress more effectively than others. Full compression ratio potentials are achieved over time and are a by-product of your back-up processes. Backup policies effect compression ratios. Some vendors quote compression rates that are not achievable without trade-offs in performance and scalability. * Find out early on whether your backup software works seamlessly with the solution. Some vendor solutions require wholesale replacement of existing backup software and/or reconfiguration of existing hardware/networks. VTLs (Virtual Tape Libraries), for example, can drive up software licensing costs and require ongoing hardware acquisition to scale. Consider the ROI from a system that leverages your existing infrastructure. * Ask for references in your area from companies that back up similar amounts of data and work through reputable solution providers with a track record of successful implementations. Deduplication enabled disk-based backup and recovery is now a field-proven solution option for backup and disaster recovery. It’s no longer emerging technology, just cutting edge technology. Beth White, Vice President of Marketing, Data Domain, Inc. Data Domain is exhibiting at Storage Expo 2006 the UK's largest event dedicated to data storage. Now in its 6th year, the show features a comprehensive FREE education programme and over 90 exhibitors at the National Hall, Olympia, London from 18 - 19 October 2006 www.storage-expo.com
•Date: 6th October 2006 • Region: World • Type: Article •Topic: IT continuity |
