Reducing Data Storage Costs Through Active Archiving

- Floyd Christofferson, Director of Storage Product Marketing, SGI

The mainstream adoption of HD, 3D, mobile and streaming services presents an archiving challenge for the digital media industry, in scaling storage and support systems cost effectively, and therefore providing sufficient capacity and speed of information retrieval required.

Even though more and more digital media files are filling up ever-larger disk silos, propelled by the proliferation of mediums, the amount of data is growing quicker than the need to access it. For the digital media sector, specific files are rarely accessed, but the key is for that access to be immediate and fast; business users and consumers want them available at all times.

For some businesses this challenge would be addressed through better data management, but translated to the digital media and production sector the challenge becomes astounding. The problem is more critical than the realm of personal preference where online media is accessible instantly, such as services like BBC iPlayer or Spotify, instead the issue is a business necessity. The business needs to have access to the full range of data at all times.

Always-on and accessible data

An active archive means data is always available in an 'online' state. In the context of an active archive, 'online' means that the data is available in an environment that is immediately and easily accessible to users, that is not drawing power or taking up unnecessary space, and one in which the data is protected for a long time.

An active archive strategy, when properly applied, significantly reduces overall storage and data management costs whilst improving efficiencies and the ability for users to access all data.

In essence, the data should live where it is most efficient. For example, inactive data, WHICH IS BLA, which has retention value can be moved into an archive tier storage that, although 'online' and visible to the user, is typically in a powered-down state using Massive Array of Idle Disks (MAID) technology that completely removes power from the array. These archives, while still available to users, can be managed with very different disaster recovery techniques that require less investment, and at a fraction of the operational costs of conventional disk-based file stores.

This is a vast contrast to a traditional archiving approach, where data often ends up residing in an off-site data tape store that required hours if not days for data retrieval.

Implementing an active archive

There are numerous tools that can simplify the implementation of an active archive strategy. These can be categorised as:

- Digital Asset Management: Leading digital asset management systems automatically index content in multiple ways as it is created and modified. Using this meta data users can search for data, and administrators can easily set policies to automatically determine which data should remain on production disk drives and which can migrate to lower cost, higher efficiency second or third tier storage.

- Hierarchical Storage Management (Tier Virtualisation): Another cost-effective technique that can aid in developing an active archive is to virtualise tiers of storage through the use of a hierarchical storage management solution. These enable multiple tiers of disk and tape to appear to users as one large aggregated volume even though the data is actually distributed across multiple storage types. The beauty of this system is that all the data appears to the user to be online in the high speed, expensive, production disk at all times. But in reality, even though the file appears to be right where the user put it in the file system, it has actually migrated to lower cost storage. This approach delivers dramatic overall cost savings without the need for users to learn and follow where their content is located.

- Low power mass storage using MAID: A MAID system is another significant tool in creating a lower cost active archive. By selectively powering down whole sections of the disk array until the data is needed MAID significantly reduces the power and cooling requirements of the data centre, much like tape libraries do, but with the added advantage of much higher performance and proactive data protection.

In conclusion

An active archive strategy requires effective planning and deployment of management tools. When implemented effectively it can considerably reduce the overall cost of managing a growing pool of digital data. Individual components can be upgraded or changed without impacting the user experience. In this scenario, scalability becomes an asset, and not a headache.