4 ms·

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 in

by chipsy 13y ago

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.

amirmc 13y ago
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).
- chipsy 13y ago
  As a customer, the thing that is most pressing and problematic is only incidentially also a security issue - it's that I have too many accounts and I don't want to have to manage all of those accounts and set them up individually every time I get a new gadget. A dream package would have a single sign-on that includes all of these things: Email File sync Identity provision (e.g. BrowserID) Quassel IRC (or equivalent or better two-tier chat/IM) Bitcoin wallet And of those, I think the most pressing is the identity provision, because it's a "too many accounts" solution that doesn't necessarily disrupt existing workflows. But adding more of those quickly turns it into a compelling offering. Initially, this could look like any other cloud-based service and even be freemium, while keeping the code open; Wordpress got the model for this pretty much exactly right(albeit the code itself was never very secure).
- metabren 13y ago
  > 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 https://arkos.io] * FreedomBox [https://www.freedomboxfoundation.org/ 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/ http://www.spacemonkey.com/
jamii 13y ago
The Locker Project (http://lockerproject.org/ http://lockerproject.org/) had a really promising approach to this - by centralising all of a users information from different services and offering a uniform api they effectively had a solution to the bootstrapping problem that plagues other social networks. Unfortunately it looks like Singly have pivoted away from that vision. I guess there's not enough demand for it to make any money.
13throwaway 13y ago
arkos.io does this.
insom 13y ago
Fedora are making a push along these lines, not so much pushing the need to go and do anything extra, but making it so that using modern Internet protocols like SIP and XMPP is as easy as running your own mail server (that is; not brain dead simple, but not particularly challenging). https://fedoraproject.org/wiki/Features/FederatedVoIP https://fedoraproject.org/wiki/Features/FederatedVoIP Work in Fedora 19 has gone into making federation work out of the box - it's always existed for SIP and XMPP but generally required extra configuration because it was possible to run your own server yet still basically be beholden to a provider, which is very much not the default with SMTP. (Of course, that's why SMTP is so horribly broken with regard to spam, so the pendulum swings both ways)

by chipsy 13y ago

amirmc 13y ago

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).

chipsy 13y ago
As a customer, the thing that is most pressing and problematic is only incidentially also a security issue - it's that I have too many accounts and I don't want to have to manage all of those accounts and set them up individually every time I get a new gadget. A dream package would have a single sign-on that includes all of these things: Email File sync Identity provision (e.g. BrowserID) Quassel IRC (or equivalent or better two-tier chat/IM) Bitcoin wallet And of those, I think the most pressing is the identity provision, because it's a "too many accounts" solution that doesn't necessarily disrupt existing workflows. But adding more of those quickly turns it into a compelling offering. Initially, this could look like any other cloud-based service and even be freemium, while keeping the code open; Wordpress got the model for this pretty much exactly right(albeit the code itself was never very secure).
metabren 13y ago
> 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 https://arkos.io] * FreedomBox [https://www.freedomboxfoundation.org/ 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/ http://www.spacemonkey.com/

jamii 13y ago

The Locker Project (http://lockerproject.org/ http://lockerproject.org/) had a really promising approach to this - by centralising all of a users information from different services and offering a uniform api they effectively had a solution to the bootstrapping problem that plagues other social networks. Unfortunately it looks like Singly have pivoted away from that vision. I guess there's not enough demand for it to make any money.

13throwaway 13y ago

arkos.io does this.

insom 13y ago

Fedora are making a push along these lines, not so much pushing the need to go and do anything extra, but making it so that using modern Internet protocols like SIP and XMPP is as easy as running your own mail server (that is; not brain dead simple, but not particularly challenging). https://fedoraproject.org/wiki/Features/FederatedVoIP https://fedoraproject.org/wiki/Features/FederatedVoIP Work in Fedora 19 has gone into making federation work out of the box - it's always existed for SIP and XMPP but generally required extra configuration because it was possible to run your own server yet still basically be beholden to a provider, which is very much not the default with SMTP. (Of course, that's why SMTP is so horribly broken with regard to spam, so the pendulum swings both ways)