Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Dropbox acquires Boxcryptor assets to bring zero-knowledge encryption to storage (techcrunch.com)
108 points by aborsy on Nov 29, 2022 | hide | past | favorite | 73 comments


Ugh. This “zero knowledge encryption” is NOT zpf. Just a per user PBE hiding a shared symmetric key.

I kept reading, trying to find the innovative way of encrypting large amount of data while not revealing the same key to all participant - some SSS variant or fast FHE or maybe a cool new PRE.

Also, all value added features besides shipping the bytes are gone when something like this is in effect. Search, preview, co-edit, metadata extraction.

I guess I’m disappointed / venting because of crushed hopes. I’m in this field and hoped something new and exciting is starting. Not trying to criticize the engineering work, which looks quite well done, just sad there isn’t new algorithmic development.


Boxcryptor was originally just a wrapper around Encfs and a Dropbox API client and probably hasn't significantly changed how they worked since they made it a fancy service. I assume that they rebuilt something like Encfs, but closed source.


What do you mean SSS, FHE, PRE? What would you like to see?


SSS: Shamir Secret Sharing

FHE: Fully Homomorphic Encryption


PRE - proxy (blind) re-encryption.

All of these are true “zero knowledge” methods that don’t require the encryption key to be shared, and therefore be granted irrevocably.


You didn’t explain what the right answer is.


Of course I didn’t. As I said, I was hoping for a new development.


Time to switch to cryptomator I guess...

https://cryptomator.org


Depending on the use-case rclone's crypt remote might suit.

https://rclone.org/crypt/


Another vote for rclone. Let's you easily turn any cloud storage service into one with E2EE encryption. And with 'rclone mount' you can get a network file system easily with decent performance and good local caching.


Also check out RCX if you're on Android.


Cryptomator works on iOS devices, rclone does not :)


Naive question: Why is this a reason to switch?


Boxcryptor clients (client software) probably will stop being updated sometime in the foreseeable future, which at least for mobile will make Boxcryptor nonviable, even in the accountless mode.

Unfortunately Cryptomator isn’t as reliable as Boxcryptor.


> Unfortunately Cryptomator isn’t as reliable as Boxcryptor.

I have not had this experience. I switched to Cryptomator because it was a much smoother and more reliable experience for my use cases. What issues did you have?


gocryptfs is another option.

https://nuetzlich.net/gocryptfs/


Love gocryptfs but it doesn't work well on Mac. securefs[1] works fine on both Linux & Mac.

[1] https://github.com/netheril96/securefs


do you know what advantages that has over rclone[1], which is much more popular?

[1] https://github.com/rclone/rclone


I guess both securefs & gocryptfs are more tailored towards mounting encrypted folders locally (i.e. they don't have any support at all for cloud storage systems, that part is left to other tools).

But I guess you can do the same with rclone given that it supports encryption, mounting and the local filesystem.

I am not sure however how they compare in simultaneous read/write access by multiple users.


Just today I was lamenting the lack of Document Provider APIs on Cryptomator Android, I'd like to use it to encrypt my Obsidian vault. Alas the issue is locked and the milestone keeps shifting.

Cryptomator has been good to me otherwise!


This really is the dealbreaker for me, I bought the android version, but it's pretty useless without the document provider


Any hints on setting up remote server? What service should be used for file uploads, that wouldn't require much maintenance and be stable?


Does cryptomator only work with Dropbox?


No. It's works with most cloud storage, either with their own client or cyber duck/mountain duck.


Any idea how these guys make money?


They take donations. I've donated to get a "key" for my copy of Cryptomator, but it doesn't do anything to the software, it's just there to turn off the thing that asks you to donate, I believe. I use it as the primary way to encrypt sensitive documents before cloud syncing them, so I figured it was worth a donation.


You have to pay for the mobile apps.


Another article

https://techcrunch.com/2022/11/29/dropbox-acquires-boxcrypto...

Dropbox announcement:

https://blog.dropbox.com/topics/company/dropbox-to-acquire-b...

Looks like Dropbox plans to provide client side encryption, although somewhat belatedly.


We've changed to the third-party article from https://www.boxcryptor.com/en/blog/post/new-chapter-with-dro....

Generally we prefer the best third-party article (if there is one) to a corporate press release: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu....


Thank you for the clarification. Will keep that in mind, next time!


> client side encryption

to business users


There are significant costs to encryption. Any deduplication is gone, losing a single "chunk" invalidates the entire file with no hope of recovery, and, more importantly, support costs are way higher - "I lost a password" becomes "I lost every file, I'll never recover them".


The non-business plans (Plus/Family) are already not cheap, especially if you only need a fraction of the storage space.


Dedeuplication is a huge cost at scale.


I'm not sure what you mean. Deduplication by hash is free and exactly how Dropbox works, as I recall.


Deduplication doesn’t work with client side encryption, at least not securely because it at the least exposes that the file already exists.


Yes, we agree.


Too bad. I cancelled my dropbox sub and moved to TrueNAS Scale + Nextcloud + B2 rclone last year. Dropbox could have me back anytime they want me.


Encryption is not for the poor.


The way I interpreted it was that it won't even apply to the paid "personal" plans.


Is encryption so special it requires an acquisition to implement? I figured it was just slightly more complicated than compression.


I think this acquisition is mostly because Boxcryptor is somewhat established in the German space and because Boxcryptor has been working together with Dropbox for a long time (since 2016).


Why to reinvent the wheel and investing significant amount of money, resources and audits when it can be brought for cheaper.


It becomes more complicated in terms of key management if you want to enable sharing of encrypted files with other accounts, like Boxcryptor allows.


Yeah I was trying to find more about ZKP but seems like it’s lord of a marketing gimmick than anything else. On a different note, still don’t understand why did Dropbox buy them,as the tech seems to be pretty basic. May be I’m missing something here.


If hashes weren't illegal, cloud storage providers could probably save millions by deduplicating pirated content.


What makes you think Dropbox doesn't deduplicate data?


Because if the data is encrypted as suggested here, then the hashes wouldn't match and the content won't be de-duplicated.


Data encryption can be decoupled from key encryption. Let's say I hash the content of a file, encrypt the file with its own hash, and then encrypt the hash with the user's password.

This way you can have deduplication of data, and only an attacker that knows the content of the file can decrypt the file, but since they already have the file, it does not matter.


Convergent encryption without a user specific key has two weaknesses:

1. You can tell that the user posses a known file. Such as a pirated movie. This is what this thread is about.

2. If the attacker knows the file except a sufficiently small secret part, they can learn that part.

https://tahoe-lafs.readthedocs.io/en/latest/convergence-secr...


I wonder if ProtonDrive is scaring them a little. The end to end encrypted file storage market was tiny to start and now it’s getting a big name player at last. Tresorit is overpriced, Boxcryptor needed the acquisition and ProtonDrive seems small still.

These products need to be available for end consumers too though not just b2B


This made me realize I haven't heard of any tech acquisition news in real businesses (meaning not crypto) in many months. MANGA isn't gobbling up competition to maintain their monopolies like they used to...


Going to make it available on Linux?


Great news. Users have been wanting this for years.


I have a dead simple backup plan that is basically free.

All my essential, must have info is stored on two 128GB microSD cards. One lives in a safe at home. The other I strap to my wrist. The cards are encrypted with BitLocker. A few trusted others have the decrypt/safe key.

The data is as safe as I am --- if not safer. If something does go bump in the night, all I have to worry about is my own personal safety. Even if I die unexpectedly in an accident, search and rescue will still likely deliver the data to my next of kin.

Let's face it --- no one is going to organize a search party if your on-line backup suddenly disappears or is being held hostage for "egress fees".


I hope you're using industrial SD cards. They are notoriously failure prone. And how often are you writing that data?


I update them a few times per month; whenever I have some new pertinent info that I feel needs to be added.

I keep two separate copies --- in case of failure.


Why though? What data?


Not OP, but I have multiple back-ups of things like mortgages, employment confirmations, what passes as letters of recommendation in Germany, educational certifications and diplomas, some other legal stuff thay moght be important and of course all documents from the comoanies I had / have. I do use USB sticks so.


If you have to ask that question, you probably don't need a backup.

Photos, last will and testament, legal documents, bank/accounting/tax info, photocopies of birth certificate/passport/diplomas, source code, login/access credentials, crypto keys, etc., etc.. All the same stuff that people like to use on-line backup for.


Aside from photos/crypto keys, all the other ones are easily replaceable within 2 weeks, at least in my not-the-US country. Source code would be only for any personal projects, as companies usually don’t let you backup their code.

And none of these make sense to have on your wrist 24/7.


So your argument is that everything can be easily replaced? If true, you don't need backup. If everyone was just like you, Dropbox probably wouldn't exist.

I run my own company. I have 20 years of source code and legal documents that are essential to my company --- among other things. Lots of time, effort and money is tied up in this info and it is not easily replaced.

And none of these make sense to have on your wrist 24/7.

You're entitled to your opinion of course but it might carry more weight if you would explain/justify instead of merely state.


Recently I’ve been chasing down old medical records. Once you cross the 7-15 year threshold from your last encounter with a healthcare organization, they often destroy the records even if they were digitally kept. Sometimes that is not altogether a good thing depending on your personal or genetic/familial medical history. And in the US at least, medical records are generally the property of the provider (organization). Restoration isn’t guaranteed.

Less than ten years after I graduated high school, the district had already migrated SIS platforms and had lost my freshman year transcripts. I have the only copy.

I’m in the same position as you with respect to having business records that I’m the custodian of, which aren’t readily or realistically restorable if I don’t make my own backup arrangements.

For anything truly important, you can’t assume someone else has your disaster recovery interests in mind.


Not your drive, not your data.

I know exactly where my data is at all times. And I know it is at least as safe as I am. Actually, it is safer since I keep 2 copies --- one stays at a fixed location and one travels with me at all times.

I spent some time contemplating the possibilities and this is a simple, easy plan that will likely survive even my death.


A SaaS feature backlog of course


How do you upload new data to your SD cards? How do you ensure that the SD card in your safe matches the one on your wrist?


Plug them into a computer using a USB adaptor.

I always update the one on my wrist as needed first. If there are lots of changes, I use XCOPY to scan/copy updated files only from the wrist card to the one kept in a safe. Otherwise, I may make the same updates to both manually.


Can I ask what you use to keep that card on your wrist? Have a similar method but yours seems more convenient.


I have a cheap digital watch with a "data vault" that I have added. This is a 3-D printed plastic device with a latching slot for holding/protecting the card.


https://www.thingiverse.com/thing:5254286

Get one of these tiny cases printed, glue it to the inside of a NATO watch strap, or you can forgo the case and see how it holds up.


That looks interesting but more complicated than mine. Mine is basically added in-line with the watch strap. The strap itself is unaltered.

The card slides into a slot in the side and latches into place. SD cards are designed with a latch key just for this purpose.


Can you share a photo?

I keep 3 (!!) RFID tags on my wrist using my Pebble 2SE and a NATO strap, they're 100% unnoticeable and invisible aside from added functionality.

I'd love to add storage for microsd to it; especially if it doesn't impact my current method


I doubt it would be of much benefit. It is black plastic (to match the watch band) so it doesn't photograph very well.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: