Five steps to a data archive strategy.

Author:Tongish, Steve
Position:Storage Expo 2008

Data centres around the world are being tasked with storing ever-greater amounts of digital content. This burgeoning storage requirement drives many data centre managers to recommend that the business increase investment in expensive IT storage resources. However, some realise that they can use long-term archive strategies to significantly limit additional investment. Instead of storing all data on expensive front line storage systems, they recognise that archive data can be migrated to more appropriate and cost-effective alternatives. This forward-thinking approach not only reins in IT budgets but also delivers compelling business benefits over the long term. Those data centre managers that believe an archiving approach to data growth is nice in theory but too complicated in practice, risk missing a huge opportunity. By thinking through five key issues, companies can begin to create a compelling archiving strategy.


Active data is data that is currently being created or used; static data is never changed and rarely accessed. Multiple studies have shown that 80% of all data stored on magnetic disk RAID systems (primary storage) is static data. The ability to move as much as 80% of data off primary storage onto secondary storage, such as optical, can slash management overheads. Magnetic disk storage is expensive to operate, protect and replace when compared to other technologies that are more appropriate for infrequently accessed archive data. The first step to any archive strategy is to separate active from static data in order to reduce the volume of data residing on primary storage.


You now have your data divided into two buckets: active and static. The next step is to assess the value of this data so it can be properly managed. You can assume that the value of active data to your business is high since it is currently being used. Determining the value of static data is more difficult since it is not all equal. The most effective approach is to create categories defined by the value that the data represents and place static data into the most appropriate category. This categorisation process allows you to define management policies over the life cycle of the data and once in place lends itself to automation of the process.


Once your static data is categorised according to its value, you need to determine how long each category should be retained. The most notable external factor that influences retention periods are...

To continue reading