"How Does a Company Stores It's 20/30 years old Data?"

Last Monday, we were asked to find the answer to the question that was asked by our MIS lecturer, Dr. Zaidi. As part of the requirements, we were required to answer the question through our respective blogs. Hence, the purpose of this current entry~ :))

Based on the readings I've done, most companies that stores 20-30 years old worth of data usually stores it in a magnetic tape. Apparently, it has been used for over 50 years and through a lot of changes, due to both the environment as well as the technological aspect, modern magnetic tape nowadays is most commonly packaged in cartridges and cassettes.

Below are an excerpt taken by http://en.wikipedia.org/wiki/Magnetic_tape_data_storage;

"For decades tape storage has offered cost and storage density advantages over many other storage technologies, such as disk storage. And for decades medium-sized and large-sized data centers have deployed both tape and disk storage to complement each other, with tape the favorite choice for tertiary and archival data storage. Storage technologies continue to advance both functionally and economically, and storage vendors compete aggressively against each other. Analysts are lining up on both sides of the "disk versus tape" argument.

As a basic comparison, mainframe-class tape drives, such as Oracle's Sun StorageTek T10000B, are priced at approximately $37,000 each, excluding tape libraries. (IBM's TS1130 is also representative of this storage class.) At any single moment in time T10000B tape drive can read and/or write to one tape cartridge which can contain up to 1TB of uncompressed data. Real world sequential data transfer speeds are high (sustained 120MB/second for the T10000B and 160MB/second for the TS1130) compared to disk. However, PC-class hard disks are priced below $200 for about 1.8TB. One mainframe-class hard disk still has a much lower price than one mainframe-class tape drive, so the economic might favor disk.

However, the key difference is that tape drives can exchange their magnetic media (the cartridges) frequently, while the magnetic media installed inside each hard disk is fixed and cannot be swapped. (The drives themselves could be moved if installed inside each hard disk is fixed and cannot be swapped. (The drives themselves could be moved if installed in swappable caddies at extra cost, with extra cost hot-swappable infrastructure.) Mainframe-class tape drives are almost always installed in robotic tape libraries which are often quite large and can hold thousand of cartridges. The StorageTek SL8500 library is one representative example (and IBM also sells tape libraries). The smallest SL8500 library holds up to 1,448 tape cartridges, for 1.4 Petabytes of online uncompressed storage. An equivalent amount of PC-class hard disk storage would be priced at $100,000 or more (only for the drives themselves) with mainframe-class disk drives priced considerably higher. Moreover, the tape library would likely deliver a high sustained sequential write speed, the media would be more portable and durable (for off-site storage), the media would meet or exceed long-term archival storage requirements (for reliable retrieval decades into the future), and the data center power and cooling requirements would be considerably lower. Tape drives also compress (and optionally encrypt) data before writing, at full rated speed, so the 1TB of "raw" storage is more typically 3TB of business data. Hard disks do not typically compress data in the drive (except perhaps when installed in virtual tape libraries), requiring server processing to compress data before recording. Also, the economics of this comparison are at least more complicated than a single spindle versus tape drive comparison.

That said, whether tape's characteristics versus disk are useful or not will depend on the particular data center and its data storage requirements. What has tended to happen in recent years is that the amount of data has grown exponentially, with both disk (especially) and tape participating in the growth. Solid state storage is now beginning to encroach on disk's previous near-monopoly in random access non-volatile data storage, while disk is pushing into tape's territory to some extent, particularly in situations where sequential data access is only a relatively small part of a particular data center's storage requirements."

Although we're approaching a greater technological era, however in terms of data storage, tape drives are still the most inexpensive way to store massive amounts of data. Although individual tapes perhaps do not reach beyond the 2TB offered by hard drive storage, however as mentioned above - tape drives are significantly less expensive, more durable and can support spanning data across multiple tapes for extremely large files. Aside from that, it also offers the most failure-resistant long-term backup solution available online, especially if it is meant to store large quantities of data.