You're viewing old version number 2. - Current version

7 min

Recently I've been fascinated by the concept of individuals having a "private cloud" service - a server that they own or rent, which has all the data that an individual could hope to have centralized, contains APIs and UI for remote access from client devices, and acts as the middleman for most transactions with the outside world.

The beauty of this is that if it were developed as a unified system, minimal tinkering would be necessary for a layman user. A preinstalled, auto-updating OS would do the heavy lifting - we're aiming for a situation where a bare minimum of configuration becomes necessary. Once configuration is done, all the user needs to know to reach all of their data is a URL and a password. The server itself can (ultimately) become the identity service for all remote accounts.

It has a straightforward progression of development and adoption, too. Create a web app that kludges together some messaging, identity, and storage solutions - the stuff needed to do this at an 80% level already exists, for the most part. Package it into an image installable on a VPS. Subsequently, make native apps for the client devices, or adapt existing ones, to match the experience of Dropbox/GTalk/etc. Last, build a hosting service or sell hardware tailored around simplifying the configuration process, providing backups, etc.

As you get further into the adoption process, more and more possibilities to reclaim data open up, since you'll hit the critical masses necessary to push new protocols into the system.
reply

amirmc 3 days ago | link

Yup, this is the kind of thing I'm working on at the moment and I'm finding that other folks are thinking along similar lines (but taking different approaches). The first application we want to build on top of the underlying 'private cloud' is email.

My viewpoint is that to do this properly, we need to think through things from the ground up with privacy and security built in, otherwise I don't see it as that much better than current systems (you'd just be getting the illusion of security/ownership/etc). It's really not the kind of system that lends itself to the mantra of 'move fast and break things'. Such software would also have to be open-source (to a large extent) so that others can verify security or people can deploy it to their own machines (I'd personally prefer a hosted version, knowing that I can migrate off if I want to).

The question is what kinds of business model can one build around such a product. Would you be willing to pay for such a service? If so, how much and what features would you expect from an early version? Would an app-store model make sense to you?

In any case, if you or anyone else is interested in finding out more about the work (esp alpha releases), feel free to drop me an email (in my profile).
reply

metabren 3 days ago | link

> from the ground up with privacy and security built in
Couldn't agree more!
> I'm finding that other folks are thinking along similar lines
> (but taking different approaches)

Not sure if this is the right place for it but does anyone have a list of all these similar projects? I know of:

ArkOS [https://arkos.io]
FreedomBox [https://www.freedomboxfoundation.org/]

Seems to me that something like this can be built faster if resources were combined in some way. I guess the tough thing here is that people seem to have different visions as to what the end product would look like. Perhaps effort would be best expended by collaborating on a solid, secure "base" image which could then be forked as needed.

---

My ideal system would:

Be open source.
Be built with a security mindset from the ground up.
Have some sort of full system/drive encryption.
Run on something like a Raspberry Pi with local storage attached.
Have one identity per-person for social use (public/private blog, private chat, private audio/video calling, private email, private sharing of photos, videos, private VRM [1])
Have the ability to spawn off anonymous identities that would not be traceable back to your IP/device for anonymous publishing/chat/email where required.
Be easy enough that I can give one to my parents, create accounts for them, show them once how to use it and then forget about it.
Have all of this backed up through some distributed, encrypted system like SpaceMonkey [2].
A possible late addition for mass adoption - the software could "trick" users into doing things they shouldn't (in a controlled way) and then after they've done this, tell them that what could have happened would have been quite serious (such as giving some . This would lead to more careful behaviour in general by all that use the software, which could only be a good thing.

Anyone know of such a system in the works? I'd love to contribute.

[1] VRM - Vendor Relationship Management. Keeps track of who you've shared your personal details with, generates a unique email for each "connection" (where connections can be you telling someone to flick you an email or you signing up to a site) in order to cut off communication at your convenience, not theirs. Sharing of personal details (such as name, birthday etc.) for sites that require it would initially be done directly, eventually moving to a system where this software IS the proxy by which third parties communicate with you (physical mail, messages, phone calls etc.). No more spam!

[2] http://www.spacemonkey.com/
reply

huntaub 3 days ago | link

Looking over your list of things in your "ideal system", I know that we are attempting to build something like that at Airdispatch (http://airdispat.ch) - feel free to give it a look, and let me know what you think (my email is in my profile).
reply

https://news.ycombinator.com/item?id=5993083

You want people to give away so much and propose what?!
> start by learning some Internet basics. HTML5 and CSS3
Isn't this a bright future? People all around saving the world by typing HTML tags and CSS rules. Let me clarify something: HTML is a DOM serialization. It's editing by hands is as stupid as typing UTF-8 byte by byte.
Our tools are weak. Consuming and publishing are disconnected. To publish information one has to convert it and push on server. Now it's copy, not original document. Modern web tries to solve this problem by making web the only place. Sure it requires resources and for sure it's centralized.
I'd like to have all hypertext goodness on local computer. Make interconnected notes, publish them by clicking button and don't worry about synchronization or file name. Interconnect with web, know that every page I'm interested in is cached locally, edit these pages in place.
It would be nice to use 320MHz 16MB RAM wireless router as always present server.
> You have no idea how to type HTML anymore Can you write HTML WYSIWYG editor? Can you write tooling to query interconnected data?
Maybe at least some users would not need to store data outside

https://news.ycombinator.com/item?id=4419152

This would be the "killer protocol" for the freedombox, if combined with some smart dyndns management.

Here is a use case scenario I am imagining. I define two servers for myself: home.me.com and cloud.me.com. Where home.me.com is a dyndns to the freedombox. Dyndyns being unreliable, if a tent msg cannot get to my home server, then the messages are sent to cloud.me.com and then pushed to home.me.com when it comes back online (think POP mail).

The facebook killer then, is a hosted service like cloud.me.com for non-tech people, but a seamless transition to the hosted at home service as soon as you buy a freedombox. This way you have the best of both worlds. Your face in the cloud, and long term storage at home.

Other app wishlist: tent to smtp and smtp to tent adapters for gmail killing

#toread
#readlater

From JR's : articles
1335 words - 8457 chars - 7 min read
created on Jul 07, 2013 at 07:59:40 pm
updated on Aug 15, 2013 at 11:17:57 pm - #
source - versions

Related articles
November 2014 links - Nov 05, 2014
Jan 28, 2015 posts to read - Jan 28, 2015
Read later - Distributed Internet Usage - July 7, 2013 - Aug 15, 2013
To Read - Nov 10, 2014
Oct 30, 2014 links to read - Oct 31, 2014
more >>