Two different, but somewhat related additions to the archive you might want to check out.
First, I was contacted earlier this week about a project to recover information off of an old Cray-1 supercomputer hard drive. Unlike, say, trying to get your old floppies to read or pulling an old mix tape off of a cassette, with something as old as a Cray-1 (a computer once called the “World’s Most Expensive Love Seat“), you don’t even have a place to really plug it in: functioning Cray-1 machines are rare as you can get, and even if you were to get the hard drives spinning up and read off of – where would you get the data off the Cray?
Researcher Chris Fenton has a thing about Cray supercomputers – he built a tiny homebrew version of one that used emulation to allow you to experience some aspect of Crays, from his desktop. So when he found himself with a 80 Megabyte CDC 9877 disk pack, which was quite a lot for the early 1970s, it wasn’t just a matter of hooking it up to USB. (Actually, we have a brochure for the behemoth you would put this disk pack into to read it.) Here’s what a nearly-the-same CDC 9987 looks like:
Ultimately, Fenton got the information off of the disk pack using a whole variety of techniques and experiments, as part of a research project this summer. He wrote a paper about the process, entitled “Digital Archeology with Drive-Independent Data Recovery: Now, With More Drive Dependence!” and it’s now mirrored here at the archive. If nothing else, be sure to browse through the paper just to see the customized stepper motor and reader he build to pull the magnetic data off the platters. And I was kind of understating things… ultimately he did hook it up to USB.
From this careful, forensic-quality magnetic scan of the drive, Fenton has produced a large image of the disk, one far larger than the data on it but allowing further experimentation and reading from the image without having to build a robot in your basement. And now, we’re offering this image on the archive. Remember, you won’t be able to pull this data down and go back to the 1970s, instantly – you should be reading up documentation of disk formats, learn about how pull information off of magnetic flux recording, and a whole other host of material and knowledge…. but hey, weekends are for having fun, right?
Even ten years ago, the idea of offering several gigabytes of something (that expands out to about 20 gigabytes of something) online was beyond crazy – that we’ve come so far in offering this much to so many people speaks how much the world has changed since the era of this disk pack.
Fenton is associated with the NYC-based hackerspace, NYC Resistor and it was their mailing list that got in contact with me to get this disk image up to the archive.
Coincidentally, this was also the week that two NYC Resistor members released a book, for free, which you might really enjoy. Bre Pettis and Astera Schneeweisz hatched a plan to make a book on hackerspaces at the end of 2008. They wanted to put it together in less than two weeks, and as people submitted photos, essays and other material, the project increased in size, more folks were brought in, and this month the end result was released for free.
Entitled “Hackerspaces: The Beginning”, this photo-filled book is available at the archive to read online or download. A worldwide view of hackerspaces throughout the world as of 2008, it also includes memories of spaces past and dreams of spaces future. It’s an excellent snapshot of a beautiful, technological world well worth browsing this weekend (and weekends to come).
So if you’re in the mood for advanced research or just to check out some great photos, the archive’s got something for you!
Pingback: Internet Archive just Recovered the Contents of a 1970s Hard Disk - The Digital Reader
Interesting bit about the Cray – thanks for the article. My father knew of him when he was competing with him for lab time at the U of M, then at CDC when Cray had already secured his status as a legend in the computing industry. My Dad always had good things to say about him.
I’ll pass the article on to him, he’ll get a kick out of it for sure.
Hehe; something like 20-yrs ago, I was gainfully employed as an operator on a Cray-1 (the magic cake slice), and I remember those particular disk packs well.
I remember the Cray engineer at our site (buy on Cray, get an engineer free) and I remember him having problems with a drive, so tried a spot of substitution; swapped the packs between two drives, and magically destroyed both drives and packs in the process.
I wish Chris would have emailed me about the service manual. I have others which were for earlier drives. It will be interesting to see how far he gets interpreting the sector format. Hopefully he can leave the setup in place until he verifies that the comparator that he built worked correctly by decoding the sector data. The description of the circuit that CDC used is quite a bit more complicated than what he used.
Also, Jason, I have several 80 mb CDC packs, and I can get a correct picture for you.
The one you show is an earlier 200 mb version with 12 instead of 5 disk platters.
Pingback: Old and gone – Michael Alan Miller
This is like Vernor Vinge’s books. In the future we’ll have whole university departments on Programming-Archaeology.
Boy, it’s been years since I’ve seen one of those! I worked for CDC’s hard disk manufacturing and assembly division (in Omaha, NE where all or almost all of CDC’s disk packs were made) for most of the Seventies and all of the Eighties. It’s entirely possible that I had a hand in the creation of that disk pack. Those where beasts compared to today’s hard drives. But they weren’t nearly so sensitive to environmental conditions. We used clean the disk pack platters with a mild soap solution, followed by rinsing with de-ionized water and a blow dry. Try that with one of today’s drives! Needless to say, clean rooms weren’t required to assemble these packs.
Also Al Kossow is correct that isn’t a 9877 disk pack pictured, if you look very carefully through the handle, you can see the label indicates it is an 873-24 disk pack. Even though the packs have 5 or 12 platters, the top and bottom platters are for protection, so there are only 3 or 10 platters that actually hold data.
Boy do I feel old now!
The recording format should be well documented (MFM, RLL, etc…), the file system layout may have been proprietary – in which case it may have to be reverse engineered in the event someone does not have a formal description or source code. Many of the older 60’s/70’s systems shipped with printouts of the operating system. I know DEC did, maybe Cray did this too. As an interesting side note, depending on the quality of the digital imaging of the magnetic surface, and software for interpretaton, even overwritten data may be extractable from these disks.
Pingback: Link Love: 9/9/2011 | The Bigger Picture
Pingback: Hard Drive Archaeology – 1970s Cray-1 hard drive imaged « adafruit industries blog
I use to Debug CDC PPU Code for these disks back in the early 70’s.
Pingback: DATA RECOVERY | | PcFixPcFix