Percloud proposal, 2018 edition

Percloud (PERmanent/PERsonal CLOUD) is the name I gave in 2013 to my own idea of “a REALLY usable, all-in-one, individual alternative to Facebook, Gmail, Flickr, Dropbox…”.

. This page explains what a percloud is, and why it is needed. The main points are also available in the slides presented at the FOSSMeet'18 conference in Calicut. If, at any moment while reading this page, you find yourself thinking “Duh, this is just the same as X”, please do read here. To know why this is still a proposal, instead, just read the last two paragraphs.

State of the Net, 2018 #

The current generation of centralized, corporate-owned digital platforms (Google,Facebook..) is a problem global enough to be a Davos topic because it provides, in a wrong but extremely convenient way, two online services that everybody needs:

  • “Personal box” service: individual online presence, communication and sharing (of text, video, pictures, calendars and other individual “data”)
  • “Group box” service: coordinated activity and communication inside groups, and sometimes offering of services by the same groups. Uber, AirBnB and similar “platforms”, down to Facebook groups, are “group boxes”

These days, there are many projects for community-owned versions of both boxes, usually under the “platform coop” label (for the record, I am a member of the Free Knowledge Institute, that is currently testing some “group boxes”). Personally, I believe that group and personal boxes should be well separated at all levels. This proposal is only about “personal boxes”.

Vision and main requirements #

  1. Person before “community”. Human beings are individuals first, and only after that “social”. In the long run, personal and group boxes can work only if they mirror this reality. Personal boxes must work, without any interruption of service, even if they are moved from any community or server to any other, or if their underlying technology changes. Othewise there is no individual freedom. Individuals must also be able to use independent group boxes through the same personal box.
  2. Persons before “platforms”. Implementing personal boxes as partitions of larger “platforms” means insisting with the same architecture that creates so many problems today. It would be MUCH more complicated than dealing with truly individual personal boxes. We should be “going where we need NO platforms”.
  3. Identities before “services”. All your personal activities, data and contact info should always remain attached to YOU, even if technologies and their providers change continuously. Having your personal blog at, and an email account as seriously limits your concrete freedom to own (and move) “YOUR” data. The way to go, instead, is to have a a permanent web domain for each identity first, and then attach all “services” and accounts to it: blogging at, email like, and so on. At the same time, it must be easy to have totally separate personal boxes (e.g. for work, family, friends, etc). “Context collapse” is wrong.
  4. Usable by EVERYBODY. Including minors (under parent supervision, of course), refugees and other people without home, broadband or reliable electricity (see PEAAS below).
  5. “Web compatible”!!! Anything that is deliberately cut off from the current Web and main social networks, is unsuited for mass adoption.
  6. Main/basic needs only. Blogging, photoblogging, microblogging.. The current diversification of digital social networks is, in large part, an artificial, induced need. The large majority of today’s social network users only needs the features below.
  7. Realism, and first things first. Since 2013 (when the percloud was first proposed) many other projects have been started to solve the same problems. Technically speaking, they are all better than the percloud. In practice, all they have achieved so far, and may achieve if history is any indication, is this. Perclouds have no ambition to be final solutions. They are more like lifebelts launched by Search and Rescue teams. They aim to be a viable route out of walled gardens, for as many people as possible, as soon as possible. THAT is what matters, not waiting till the perfect solution is ready.

I sincerely believe that any proposal of personal boxes that ignores these requirements has very little chances to make any meaningful and durable difference, no matter how cool and advanced it is. Anything that feels “so underpopulated that gives no incentive to go back”, because has no built-in way to stay in touch with one’s Facebook “friends” will fail, or come too late to make a difference anyway. Ditto for anything so radical to be incompatible with the current Web, or is harder to get than a few taps on a smartphone.

Governance? #

Which governance? With really individual “personal boxes” there is NO governance to worry about. Percloud puts “governance” back where it should have stayed: plain old laws and full individual responsibility. If the place where Joe Someguy writes or does something illegal is his own blog at, the authorities may do what they must, without begging some corporation as it happens today. But, for the same reason, automated mass censorship like this or this becomes much harder to achieve, for both corporations and governments.

Features and architecture #

A percloud only supports these services:

  • “content hosting”:
  • host articles, images, audio, bookmarks, calendars, any other file of the percloud owner
  • publishing and presentation (e.g. schedule posts, make image galleries…)
  • access control, at file level
  • discussions about each file
  • notify personal contacts of new content, or discussions
  • announcing new content on Facebook or Twitter accounts
  • UNIFIED “news” interface. Here, “news” means everything from new articles on magazines websites or other perclouds to new email, Tweets, Facebook status updates… A percloud owner remains connected to the whole world, without filters, in ONE “window”
  • collection of all “news” directly from their sources
  • presentation of all “news” in ONE window
  • communication: email, mailing list manager, instant messaging
  • administration (and automatic backups!): well separated from daily usage and assignable to a trusted party

What? Why is “X” missing??

Because of all the requirements above, of course. The percloud needs to stay in the lower right corner of this diagram:


In my opinion, in order to satisfy all the assumptions above, a percloud must be a Gnu/linux distro, or container, that:

  • is pre-configured with the right combination of Free Software
  • supports the main CPU families, from ARM to x86_64
  • has an installer that also attaches to it a domain name
  • is heavily optimized for what will (must) be 99.999…% of its first installation, that is PEAAS

A software bundle like that is, by definition, self-installable on any hardware that runs Linux, from home computers to the datacenter of any ordinary hosting provider. It can be moved from any computer to any other without losses, because from the outside everything remains “attached” to “”, remember? New features can be added without interfering with the others. Finally, once all one’s data inside such a box, in open formats, it is way easier to migrate them to other solutions, than directly from, e.g., Facebook.

PEAAS: Percloud as a service #

“PErcloud As A Service” means getting your percloud created, hosted and managed by any web hosting provider, in their datacenter. It’s the same thing that already happens today when, for example, setting up a blog at a few clicks, and you’re done, even from “Free as in Freedom” smartphones.

That is the only way to make perclouds really usable by the people who need them the most: those who don’t want to, or simply cannot, install and manage anything by themselves. Consequently, besides having a well separate “control panel”, perclouds must also include hooks for automated installation, configuration and maintenance of many perclouds in one datacenter.

Software choices #

“Anything, as long as it is black”. If you disagree with the main vision, why argue on the components? This said, and just as a starting point, many components may still be the same I listed in 2017. Or those of the closest thing to the percloud that I know right now, Indienet. Or scuttlebutt. Then again, please do read here before concluding that percloud “is just like X, therefore is unnecessary”.

Development and deployment #

I believe that, sofware-wise, a percloud is simpler than all the other projects I know of in this space. In any case, its worldwide adoption would need no coordination or huge infrastructures. You don’t need anybody’s permission or coordination to set up your own, worldwide-interoperable, email server. The percloud is the same. If it already existed, one million people could set up their own instance tomorrow, each by herself, no problem. Any community with enough money to offer PEAAS to all its members, could do it in the same way.

In spite of this, developing and testing both the actual software, and its sustainability in PEAAS mode is no simple task. There should also be an app, possibly bundled with projects like Eelo, to let people create a PEAAS percloud, and its associated domain name, from any smartphone.

Besides that, the biggest feats are, I believe, notifications among perclouds, a homogeneous and really user-friendly interface, and tuning it for real world PEAAS usage. The latter part would also need the field test/pilot I described last year.

Who MAKES it, and who PAYS for it? #

Who MAKES it? I already explained in FULL detail why this is still only a proposal and, specifically, why I keep making it. But here is the short version:

whoever likes this idea, and has the time and skills to realize it, please just do it, giving credit where credit is due.

Why? Because, as much as I’d like to work on this, and I’m sure I could do a good job, I could never do this for free, and it could only be a (paid!) team work anyway. By myself, I don’t even have time (or skills) to seriously search for funding in any way.

“Who pays for it?" is actually two, equally important questions:

Question 1: Who would pay for development and maintenance of all this software?

A1: whoever cares for a really decentralized, multi-cultural Internet that puts privacy and other individual rights at the center: EU, UNO, NGI, Open Society, Public Administrations, who else?

A2: whoever would make money from offering PEAAS. The main such category would surely be hosting providers.

Question 2: Who would pay for the HOSTING of perclouds?

A: Using a percloud via PEAAS (=99.99…% of cases) cannot be free. A (very?) few dollars per month, yes, but not free. I propose different answers:

  • those who can afford to pay the hosting, will do it. The pilot I suggest has also the purpose to know how much PEAAS hosting should cost
  • there are many organizations that already, implicitly have, or otherwise should have, providing perclouds to their own “members” as part of their mission: governments, local administrations, schools, NGOs helping disadvantaged groups… All these organizations may (and again: in my humble opinion, often should) pay (part of) the hosting costs for their members, maybe only for one or two years. After that, those members could migrate their percloud to any other PEAAS provider

If you read this far… #