One might quibble with this contention. Maybe some information has value in terms of big data and similar analytics. Still, the basic question to consider is whether you can dispose of unnecessary data or, at the very least, move it to more cost-effective storage. The answer is not easy; in fact, the problem is very complicated.
Meanwhile, the cost of infrequently (if ever) accessed content devours a significant portion of the storage budget each year. Moreover, pundits promise a continued deluge of new data with little if any guarantee of enhanced value or use. How can organizations afford that sort of potential waste with IT budgets so tight?
Now, a good prescription to deal with the problem is information governance, plus analytics software to help identify data value and perform other tasks, such as data disposition. While this is what should be done, such a process requires time, organizational commitment and cost. A more immediate approach is to move data to a more cost-effective storage medium than hard disk drives, such as tape.
Yes, this treats the symptoms rather than the disease. However, tape provides a more cost-effective medium by storing data in the form of an active archive using a file-system-based strategy. The active archive can be composed of a mixture of useful and not useful data without having to distinguish which is which. Moreover, the data is likely to be able to tolerate the higher performance latencies (slowness of access/delivery) that is common in tape.
Why not leave the data on hard disk? The answer is that tape for active archiving is well over an order of magnitude more cost effective than hard disk. For more, see Clipper Group reports at the LTO Consortium.
Why hasn't tape been used more extensively? Although there have been successful tape solutions, most are proprietary technologies. Only in the last couple of years has a non-proprietary technology emerged to make tape feasible as a native tier of the storage hierarchy that also supports necessary coupling and integration with other technologies.
This base technology is called LTFS (Linear Tape File System), which makes writing to tape as easy as if it were disk using a file system. IBM provided LTFS to the LTO Consortium (whose leading technology providers are HP, IBM and Quantum) for members to download. It also contributed LTFS to the Storage Networking Industry Association (SNIA) for standardization. Any vendor that uses or offers LTO technology has the option to use LTFS.
IBM first offered LTFS in conjunction with LTO as a single drive edition in 2010. The primary limitation was scaling, so IBM introduced a library edition in 2011 that presented a tape library as a single mount point for a file system. Later that same year, IBM introduced the LTFS Storage Manager. Among its capabilities are an intelligent copy manager to automate movement of files in and out of LTFS-managed tape storage and a metadata database to tag content and search.
These capabilities have made using an LTFS-enabled tape library much easier, but IBM felt that there was more to be done, including the policy-based movement of data between disk and tape. This means that business rules could be defined and put in policies for the movement of data to different pools of tape storage that can be managed appropriately for either preservation or disposal.
To help correct that situation, IBM announced the integration of General Parallel File System (GPFS) with LTFS (as the LTFS Enterprise Edition) during its EDGE 2013 conference in Las Vegas. By moving data from hard disks to tape, customers may alleviate some of the cost pressures that arise from storage overburdened with often unnecessary and useless data.
Tape can now be a true member of the storage hierarchy using LTFS, as well as technologies that enable tape to work in conjunction with it, cost effectively in an enterprise-scale IT environment. An active archive can store data that may need retrieval at some point, such as rich media or video surveillance data.
But it can also serve to store tons of presumably no-value data more cost effectively until owners reach a decision on defensively disposing of select pools of information. That serious money savings also means that you can use freed up HDDs for more important uses, such as the incoming rush of big data streams.
Yes, there are a lot of exciting developments that IT needs to work on, but don't ignore the basics, especially if they have relevance to your business.
IBM is a client of David Hill and the Mesabi Group.