History 2005: Case fo CAS (Editorial)

It’s the hottest jewel in the crown of computer storage, the stuff that both vendors’ and users’ dreams are made of.

Gone are the days when you search around for hard copy documents in dusty file cabinets, gone too are the days of waiting minutes on end for data stored until now on sequential magnetic tape cartridges or optical disks stacked away in jukeboxes.

Now, in only a fews, information can be located and extracted from tens or even hundreds of terabytes, and pulled up on the screen, and furthermore, in complete security, fulfilling government and regulatory compliance requirements, at least according to its supporters.

The new products, called CAS (Content Addressed Storage) are giant boxes stuffed with low-cost ATA or SATA HDDs protected in HAID with certain peculiarities that set them apart from traditional disk arrays. When a file is stored this way, it is first compared with those already saved, in order to save space by deleting duplicates (for example, in email attachments). In addition, a special algorithm provides, for each incoming file, a unique address or signature, which is calculated based on the content. The user can prohibit the deletion of certain saved data, then erase everything after a given date.

Basically, it’s the ideal electronic archival system.

Of course, CAS is also very costly. There’s nothing for under €150,000 for 10TB, a price that will triple once you include the necessary software to run it.

EMC’s Centera is the best known, but there’s also a choice of NetApp’s NearStore 200, Archivas’ArC, IBM’s TotalStorage DR550, Hewlett-Packard’s RISS, Sun’s Compliance and Content Management Solution or StorageTek’s Lifecycle Fixed Content Manager 100 (available in the US only).

Even so, should we fully trust a CAS system to store, worry-free, the entire legacy of an enterprise, knowing with certainty that any piece of data can be found intact and usable in the future?

To begin with, the use of magnetic drives may worry some people, since they are erasable media whose reliability is not eternal, even with the highest levels of RAID.

Manufacturers have also evoked the simulation of WORM disks, thanks to an erasure protection system implemented via the software. But you can’t really compare a true WORM medium, such as CD-R, DVD-R, Plasmon’s UDO or Sony’s PDD, where the laser burns actual holes that are impossible to plug later on, with a disk where the magnetic field can always be manipulated physically, despite the software protections in place.

What’s more, everything about the CAS system is proprietary: the APIs that grant access, the indexing systems and the unique signature of the data, not to mention the drives, which can only be supplied by the CAS vendor, not from any old disk drive maker.

Worse yet, encryption specialists have just proven that it’s possible to crack the MD5 hashing algorithm employed by Archivas and EMC for the digital signature.

Another question: what is the legal value of documents saved on magnetic media, even if these have been “worm-ized” with simple software protection? There is almost as much legislation on the subject as there are countries in the world, and the legal wording remains fairly vague, always specifying what would be most desirable, without actually preferring one technical approach over another, or one medium over another.

“Centera has legal value,” asserted an attorney (paid by EMC), expressing his opinion on the subject.

Don’t assume that a company that sells CAS will guarantee vs. data loss or legal liability in an eventual lawsuit where the judge deems the data stored on that particular system is not sufficiently probative. From time immemorial, with computer systems, the warranty covers the hardware, never data loss or legal fees and fines from lawsuits brought about by system failure.

CAS does seem well-suited to the archiving of certain information, such as email, certain accounting documents, corporate documentation, etc., but we wouldn’t recommend storing the plans for a nuclear reactor or notarized documents, unless an original (hard copy) is also filed away somewhere, bringing us back to the starting point, with a parallel system of paper archives.

Certain documents need to be maintained for several dozen years, depending on their content. In 20 years, where exactly will the Centera maintenance technician be, where will we find that replacement for a defective part, will EMC even still exist?

The only media that have really proven their durability are engraved stone and written paper. Microfilm, which is almost impossible to counterfeit, is also fairly reliable, since it will never require much more than a lamp and a magnifying glass to read it.

What about optical WORM, the real thing? Remember those hardy old 12-inch optical disks (ATG, Philips, Sony)? They’re gone now. The CDR, more and more edged out today by the DVD-R, may end up like the old floppy disk – forgotten.

Do we need to return to the stone age, or the glory days of the Gutenberg printing press, or even re-outfit all our offices with microfiche readers? Of course not, we live in the times. Which doesn’t mean, of course, that in a desire to access, as quickly as possible and at all costs, the data we depend on, we should rush forward, and in our haste, put our very heritage, corporate or otherwise, at risk.

J-J.M.

This article is an abstract of news published on issue 206 on March 2005 from the former paper version of Computer Data Storage Newsletter.