The Everything Drive | Hacker News Reader

The Everything Drive

slate.com

15 pointsvmyy9911 comments

A bit outlandish at first sight, but a very interesting idea. RAID over cloud storage services! Call it RAICS. You could even build some actual redundancy into it, just in case one of them goes down. Also add client-side encryption, and I'm in!

But honestly,

> if you’re paying anyone for online storage, you’re a chump.

No. If you're not paying anyone, you're everyone's product.

The problem with such a service is that your typical client would be the sort of person that doesn't pay for anything. An interesting technical challenge but hardly a viable business opportunity.

I don't agree; a lot of those people with lots of data are not necessarily going to be technical enough to, say, jailbreak an iPhone and install pirated apps, so it could work as a iOS app. Or it could be implemented by cloud-enabled OSes with different revenue models (desktop Linux, Android, etc).

The problem in using multiple services is split between pricing, limitations, and to a certain extent attack surface area.

If you start using lesser and lesser known services, the greater the potential threat of odd limitations (SkyDrive's 2gb file size limit, for instance), and especially apparent, a greater surface area for compromising someone's stuff. Trusting 7 companies with data, each secured with a (hopefully different) password is just too big of a risk to practically take. I'm slowly switching to Google Drive for most of my important stuff (I have my eclipse workspace in Dropbox) as a result of two factor and a proven track record of not being compromised. I keep raising the point, but Dropbox has already made this mistake once (http://news.cnet.com/8301-31921_3-20072755-281/dropbox-confi...). Though they're very likely more secure as a result, I'd be hesitant to trust any company lesser known than them for security.

When it comes down to it, a lot more goes into choosing a space provider than just the amounts offered free - making a decision on that alone is bound to lead down troublesome paths.

I don't see why you'd trust any company with your unencrypted data. Use encrypted disk images.

TrueCrypt is free and cross-platform: http://www.truecrypt.org/

If you're only on Mac, encrypted disk images are very easy to make and use: http://matthew.mceachen.us/blog/free-easy-encrypted-storage-...

Yes, you can put a TrueCrypt volume in your Dropbox, and Dropbox is even smart enough to upload only those bits that you change. But it means you lose the benefits of syncing individual files between multiple devices. Good luck trying to mount the same TrueCrypt volume in two devices at the same time. Instant conflict!

The same problem happens with every encryption solution that uses a large container. The only viable solution is filesystem-based encryption schemes that work with individual files, like EncFS or eCryptFS.

By the way, SpiderOak has client-side encryption and two-factor authentication. But it's slow as molasses.

Seeing as dropbox stores the file local, you will not have any problem with mounting it simultaneously. You couldn't merge the changes, though.

The bigger problem would be to access it e.g. on your iphone/from the web. For that reason, the webservice should be able to serve an unencrypted part if you enable that functionality. upon providing the passphrase, of course and waiting for the decryption. that could be a special feature which costs more, but would definitely add something special, which others do not provide as of now.

In most cases, merely mounting a filesystem causes a few bits to change, unless you take special care with your mount options. I'm not even sure if you can tweak mount options with TrueCrypt. So I still think that per-file encryption systems like EncFS work better with cloud storage than container-based systems like TrueCrypt.

Putting encrypted files into a "cloud" service seems more likely to lead to a situation where the file owner is placed in a windowless, soundproofed room somewhere and offered a choice between decrypting those files or being sodomized with a mop handle by an employee of a nameless government agency than if the file owner kept those files on their own devices rather than advertising their existence to the world at large.

The idea is not new (and certainly I'm not the only one using this since years - there must be hundreds of thousands doing it if not soon millions of users).

I guess integration of storage aggregation might soon be seen in various apps (if ToS of the storage services allow) as more and more apps now can make use of online storage and services like box.com, dropbox and the many others that started about 4-5 years ago have made normal computer users familiar with these concepts.

As mentioned again in one of the links within the article (http://pandodaily.com/2012/02/26/steve-jobs-was-right-dropbo... - was previously discussed here on HN) providing online storage is a feature and not a product.

Like with many other things before, pure-play online storage is becoming / always was a commodity service. I guess providers will continue to compete around add-on services. That can only be good for consumers of these services.

Could try SMEstorage. I think they have an "aggregation" concept like this.

I think this article misses the point of the these new cloud solutions though. I'm not a fan of aggregation.

The next big stage of cloud storage is integration. Integration into 3rd party apps. Integration deep into devices (iCloud for iOS and Google Drive for Android). Coupling them together will miss the biggest advantage of cloud storage.

Reply on news.ycombinator.com