Quoting Computer Weekly

File archiving is a valuable tool for any storage organisation that needs to free up storage capacity -- shifting unused data off of active storage systems. For example, you may not know if you'll need a piece of data in five years, but you don't know if it's safe (appropriate) to delete it either, so you place the data into a long-term storage system where it's accessible if needed. That long-term storage is the archive. Design documents, software builds and images are just a few of the data types that are often relegated to archives.

Of course, many companies save data to comply with government, industry or self-imposed regulations. There's a great deal of concern with established regulations, like the Sarbanes-Oxley Act (SOX) or the Health Insurance Portability and Accountability Act (HIPAA), but the recent changes to the Federal Rules for Civil Procedure (FRCP) state that anyone possibly subject to a future court proceeding has to be able to produce the required information. This impacts a much broader audience than SOX or other compliance regulations because anyone who might be involved in litigation must take appropriate steps to protect and preserve their data.

Still, if you determine that a piece of data has no possible future value or bearing on future litigation, there is no reason to archive it -- you can delete it. Data that can be reproduced or regenerated may not need to be archived. Another school of thought takes a more conservative approach and keeps every piece of data.

How does file archiving differ from file virtualisation?

File archiving is about moving inactive data. By comparison, virtualisation is about abstraction, for example, insulating applications from storage. With the recent buzz about file virtualisation, some vendors try to describe file archiving as a form of file virtualisation. Others argue that file virtualisation is about global namespaces and global file systems, clusters, heterogeneous file replication, etc.

Keep in mind that virtualisation is typically associated with virtualising the file system and allowing you to perform different operations. So, migrating inactive data through file virtualisation (even part of an automated tiered storage environment) could potentially overlap the definition of file archiving, though file virtualisation offers many more potential capabilities.

What are some policy managers for implementing file archiving and where do these policy managers exists?

Policy managers take your business rules about the information being managed, then implement those rules to move, store and delete information in the enterprise. Trying to find a policy manager can be tricky because they are not necessarily discrete hardware or software products. It's really a feature that is built into numerous solutions. Generally speaking, policy managers will run in software on host servers or on appliances within the network or on storage subsystems.

Typically, a file archiving package will have some policy manager, but some policy managers can also talk externally with other policy managers or data mover (file migration) products to ensure a uniform execution of business rules across storage systems.

What is the best medium and type of device for file archiving?

There are lots of different thoughts. I'm not convinced that there's a single silver bullet that's the best all-around medium for all types of archiving or environments. But there are some very good mediums, including disk drives, which are both fixed and removable. Some drives can even "sleep" where they spin down for prolonged periods to save power and wear on the drive mechanism. Removable disk drives can be sent off site like tape cartridges. Even traditional tape is still employed for archiving. Optical WORM drives are also used, though these occupy more of a storage niche.

The choice of medium for file archiving really depends on your preferences and business needs, so there are many options to choose from, each with varying levels of capacity, performance, data preservation capability, reliability and cost effectiveness (not necessarily the cheapest). Consider your requirements, including the features that you absolutely must have, and then contemplate any value-added features, such as compression, data deduplication or replication.

Read the original, here.