this post was submitted on 07 Apr 2025
148 points (95.7% liked)
Technology
69046 readers
3236 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related news or articles.
- Be excellent to each other!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
- Check for duplicates before posting, duplicates may be removed
- Accounts 7 days and younger will have their posts automatically removed.
Approved Bots
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Massive deduplication across all accounts on all servers of image, audio, and video data would theoretically be possible, but ain't gonna happen. Or we could just discourage people from posting cat videos and bad memes (even less likely to happen).
Deduplication is trivial when applied at the block level, as long as the data is not encrypted, or is encrypted at rest by the storage system.
If the storage all belongs to one machine, yes. If it's spread across multiple machines with similar setups that share a LAN, then you need to put in a little thought to make sure that there's only one copy for all machines, but it's still doable.
In this case, we're talking millions of machines with different owners, OSs, network security setups, etc. that are only connected across the Internet. The logistics are enough to make a hardened sysadmin blanch.
I would argue that duplication of content is a feature, not a bug. It adds resilience, and is explicitly built into systems like CDNs, git, and blockchain (yes I know, blockchains suck at being useful, but nevertheless the point is that duplication of data is intentional and serves a purpose).
Git only duplicates blobs; textual content is generally stored as deltas (look at git_repack for more details). And it's bad practice to version-control blobs: the more correct approach is to control the source from which the blob is generated.
CDNs don't all work alike so it's impossible to generalize. I won't comment on blockchain, since in my work as a developer and architect, I've never encountered a valid use case for it.
You're missing the forest for the tree here.
Given identical client setups, two clones of a git repo are identical. That's duplication, and it's an intentional feature to allow concurrent development.
A CDN works by replicating content in various locations. Anycast is then used to deliver the content from any one of those locations, which couldn't be done reliably without content duplication.
Blockchains work by checking new blocks against previous blocks. In order to fully guarantee the validity of a block you need to guarantee every block, going back to the beginning of the chain. This is why each root node on a chain needs a full local copy of it. Duplication.
My point is that we have a lot of processes that rely on full or partial duplication of data, for several purposes: concurrency, faster content delivery, verification, etc. Duplicated data is a feature, not a bug.
Technically git is a blockchain
If the data has value, then yes, duplication is a good thing up to a point. The thesis is that only 10% of the data has value, though, and therefore duplicating the other 90% is a waste of resources.
The real problem is figuring out which 10% of the data has value, which may be more obvious in some cases than others.