How do I use file redundancy?

Discuss anything related to portable freeware here.
Post Reply
Message
Author
User avatar
webfork
Posts: 10821
Joined: Wed Apr 11, 2007 8:06 pm
Location: US, Texas
Contact:

How do I use file redundancy?

#1 Post by webfork »

So there's a class of software for creating reliable/redundant data called Parity Archives that you may be familiar with. The idea is that you can use it to shore up potential problems with data that may occur in the future or during transfer. The beauty of this process is that you don't have to find another copy of the data or re-transmit, and you only need a file that's a percentage of the original to fix problems.

You can read more about it on Wikipedia, but one thing I noted in the article is that it's "mature for widespread use" (referring to the PAR2 format), but is really being used? I'm reasonably technical, take great care with my backups, but even after spending some time reading about it, I'm still unclear on it's use. This might be a real life-saver for certain users/groups -- can we make some recommendations?

Anyway, I tried to put together a listing of my assumptions about it and open questions (below). For our purposes, I'm focusing on Multipar just because it appears to be in active use and development.

---

Who would use this? (use cases)
  • Backups - either long term or on on unreliable media (e.g. an older hard drive). Also see Midas' post on bit rot.
  • Poor connections, or where re-sending data would be expensive or problematic .
  • Unreliable file servers - specifically newsgroups, where PAR files were originally popularized.

Associated tools - if you're trying to copy a file that's been damaged and keeps giving an error message:

Open questions
  • Is there a cogent recommendation on redundancy? I'm assuming (in the case of Multipar) 100% redundancy is 100% reliable but is that reasonable? Can we recommend something based on your situation?
  • A way to do command-line operations to batch add it to your file or files?
---

Did I miss anything? Is the wrong solution to the problem? Maybe there a better method than PAR.

---

EDIT:

I realize this is a broad question so I'll also add a specific one: when I was drafting the entry, I wrote up "10% of the original for already safe methods, 50% for frequent failures", but is that accurate?

Post Reply