• Please review our updated Terms and Rules here

1987-1989 Backup Technology/Practices?

charnitz

Experienced Member
Joined
Nov 17, 2020
Messages
55
Location
North America
What were the best backup technologies and practices during the 1987-1989 period for non-networked computers?

A few use cases (all on a 286 with an internal hard drive to stay within this forum's hardware period):

An author writing a book who wants to ensure the manuscript is not lost before it goes to the publisher.
A person keeping a journal to preserve and access "indefinitely".
Someone keeping an important database for a small business.

Was it just keep backups on one or more floppies? Tape drives? Something else? What were the accepted practices around redundancy so that the backup(s) weren't found to be useless right after the internal hard drive failed?
 
The database would be a prime reason for the use of backup floppy software if not tape. Relatively large data set that can compress nicely means the backup set might only take 10 high density floppies. Add a second hard drive or go for one of the much higher capacity drives compared to the typical 20 to 30 MB drives and tape looks very appealing. Keeping a 50 floppy disk backup in order gets a bit complicated.

Backups should be verified during the operation. Rotating backups were ideal but few at home kept the preferred 3 complete backups. I know of some that purchased an AT and upgraded the hard drive but kept the old slow limited capacity hard drive to backup some critical files. Losing two drives at the same time implies that a more significant business failure happened. Not too many home users or small businesses had a backup system that could be recovered after the building burned down.
 
In my former experiences and mostly as a retired federal employee, common desktop backups in that era were for data only and mostly done on floppies. The LAN systems were quite different from the present, and usually involved a mini or mainframe setup using tape. Sensitive data went out to HQ via a T1 to a massive data center and was preserved on tape.
 
There were tape backups almost from the very first; e.g., Ampex Megastore; VHS adapters, etc. I was using DC600 QIC carts back then.
There was at least one 5160-era backup utility (FastBack?) that used floppies, but used their own filesystem and format to use every bit of the floppy.
 
http://www.oldskool.org/guides/dosbackupshootout is a more recent overview of floppy backup programs. It is easier to find than the comparative reviews from PC Magazine.

Tape had a fairly high upfront cost so was uncommon for stand alone systems. My experience with QIC was that backups failed verification often enough that I usually made sure to grab two tapes to be ready to repeat the process. At least, tape backups didn't require operator intervention so I could do other things before checking the verification status.
 
Well, there are QIC drives and QIC drives. Consumer-level ones were terrible. QIC-36 and QIC-02 interface drives were good enough--and reliable enough for commercial use. The QIC "floppy tape" drives were pretty bad.
 
There were tape backups almost from the very first; e.g., Ampex Megastore; VHS adapters, etc. I was using DC600 QIC carts back then.
There was at least one 5160-era backup utility (FastBack?) that used floppies, but used their own filesystem and format to use every bit of the floppy.

Fastback was just that. Fast. I have a copy around here somewhere. IIRC, PC-tools also offered a backup utility. Then there's always MS-DOS backup & restore. Don't recall when they started offering that.

I imagine an old Colorado 120 tape drive could be obtained for not too much money now. They were bloody expensive back in the day.
 
Backup floppies of important data

What about floppy reliability? I have heard the following:

  • Floppy disks were very unreliable.
  • Floppy disks produced during mainstream production were reliable, but floppy disks produced at the end of mainstream production were not.
  • Floppy disks were reliable in the short term, but not in the long term (regarding the second use case of someone wanting to preserve data indefinitely)?

If you agree there was a shift in reliability, when did it take place? After 1987-1989?


The database would be a prime reason for the use of backup floppy software if not tape. Relatively large data set that can compress nicely means the backup set might only take 10 high density floppies. Add a second hard drive or go for one of the much higher capacity drives compared to the typical 20 to 30 MB drives and tape looks very appealing. Keeping a 50 floppy disk backup in order gets a bit complicated.

Backups should be verified during the operation. Rotating backups were ideal but few at home kept the preferred 3 complete backups. I know of some that purchased an AT and upgraded the hard drive but kept the old slow limited capacity hard drive to backup some critical files. Losing two drives at the same time implies that a more significant business failure happened. Not too many home users or small businesses had a backup system that could be recovered after the building burned down.

Thank you for the detailed description. Could you please describe how 3 complete backups were preferred and used in practice? By backups verified during operation, did this mean actually restoring the operational data from backup and using the restored data operationally? Also, if there was any failure rate in floppy disks (see above), wouldn't a backup set increase the odds of a complete data loss if any one of the floppies failed in the set?
 
Floppy disk reliability isn't a factor with the 286 systems mentioned. The double sided 5.25" disk problems had largely been resolved by the time of the release of the IBM PC. Earlier difficulties with disks and controllers had led a number of systems to have the option of backing up to cassette tape but that doesn't apply to the IBM PC compatibles. More recent disk production may have been problematical but no one was going to do a full system backup on floppies after 2000. A 128GB hard drive would require about 80,000 floppy disks for backup which would fill a 10' by 10' room floor to ceiling and weigh as much as a small car. Not a practical solution.

The quality of disks depended on where one bought. Preformatted disks from major companies had a 99%+ success rate. The company I worked for in the 90s bought the cheapest disks possible and had a roughly 50% failure rate. Had the purchaser been instead for whichever major brand was on sale, the price per working disk was lower than the bulk purchase and bulk purchases of quality disks would have saved a lot of money not to mention developer aggravation. Good backup software would kick out a failed disk and accept a replacement at that point of the backup cycle and other than the time wasted in waiting for the disks to be tested the entire set would complete successfully. Some backup software tried to gain a little speed by not testing and also lacked good verification tools but a backup that can't be restored is useless.

The 3 backup cycle was a concession to how systems would run. With tape, a backup with typically be run over the weekend. Next week, plug in a second tape to get a second backup. Third week, third tape, third backup. Then, because of how much tape costs and the fact that falling back to very old backup isn't much use, the first backup tape would be overwritten with a new backup. Repeat the cycle and replace worn out tapes. At any time, there would be at least two backups available to recover from. If both failed, the backup administrator would be looking for alternate employment. It is possible to do the same with floppies.

Floppy failure should not be much of a problem. If a floppy was bad at the start, backup software rejects it before writing to it. Most disks should have data integrity for years so a set could be restored. It is possible that a given disk might fail between writing and restore but some of the better floppy disk backup software would restore the rest of the disk set and flag the damaged files that couldn't be restored. Maybe a second backup or install disks for programs could be used to fill in the files that failed to restore. No backup procedure can ensure 100% success but it was possible to get a very high success rate affordably with just a little care.
 
Floppies for sure.

Outside of actual programs, especially on a personal scale, data was NOT that big. A single 1.2MB floppy could hold a LOT of "book". 200,000 "words".

Tape backup certainly existed, but it wasn't "cheap" (honestly nothing was cheap back then). Floppies were ubiquitous and "good enough" for a wide array of scenarios.

For a file server, supporting several users or a database, then, yea, a tape device is certainly worth while.

But more most personal uses, floppies for the win.
 
The WORM discs were very expensive at first, more than $100. Law firms, who needed a permanent unchanging archive, loved them. Everyone else wanted something cheaper.

A 1990 publication comparing various archiving measures can be found at https://pubs.usgs.gov/of/1990/0575/report.pdf Some of the prices (like for floppies) seem slightly skewed to justify their choice of WORM drive. It also does not consider the major problem with WORM drives that media wound up quickly becoming orphaned as new incompatible drives were introduced.
 
Tell me about it--I've got a stack of the things that I'm still trying to puzzle out! :) Also, WORM was not fast; about the speed of a somewhat souped-up floppy.
 
Back
Top