Status November 2019

2 weeks ago, Tim came to visit me, and we discussed a lot. A lot has changed since the last status update. And we wanted to share what is happening behind the screen :slight_smile:

Current financial situation

The current financial situation got better, we’ll share some news around this in the coming months, but we finally have enough to pay Tim full time and pay a freelancer from time to time to help on the infrastructure.

Infrastructure

We are still rationalizing our infrastructure, we’ll contact you on a case by case basis to migrate you to our new infrastructure. This will start to happen in December.

Plan for the next 3 months: αβμ

μ - the meta

We want to clarify what is IndieHosters and what it is offering.

There was always a tension in IndieHosters between the fact that we want you to be a member that decides of everything, and we want to be in a worker owned cooperative, where we, workers, are the only one deciding.
There was also a tension between the not for profit, and the fact that well, we want to pay our rent with our services.
We want to make hosting cheaper so that more people can quit google, and provide a nice service.
We want to be able to support people.

As you can feel, not everything is compatible. And I think we start to have an idea that we feel comfortable sharing.

As a first point, we have different users. We have people that would like to be member of an association, that like this idea, and others that are more client.

We came up with the idea of a virtuous triangle, here are the 3 angles that help each other:

  • IndieHosters - the association
  • oxalis - the cooperative that Tim is part of
  • libre.sh - the free software project that powers our infrastructure

The association

The association already exists. And the idea is to transition most of our users there.
We’ll keep a fair price, the support will officially be only on the forum - community support.
And this is where our pro-bono activity will also be.
We want you to be more involved there, and we’ll contact you in the coming months to see how we do. But if you already have ideas, feel free to expose them here, or privately by email.
The 1st of February, it is our 5th anniversary, and we’ll run a crowdfunding to get more members there too. And we want to release a new shiny platform (more on that later).
The draft of the presentation of the association is already here. Feel free to edit or ask questions here.

About your debt :slight_smile: the bills we never sent you :slight_smile: We’ll send you the bills in January to prefill the crowdfunding. We’ll send you a reminder, but don’t worry if you don’t receive the bill, we’ll keep running your service, and renew your domains (until the crowdfunding hehe :slight_smile: ).

We feel it is our DNA, and we want to keep doing it. It is also a great place to experiment new things with you, and see what works and what doesn’t.

The cooperative

If you need a profesionnal service (High Availability, SLA, phone support, consulting, customization, training…) this is where you go. We leave the name to the association, the split is not clear yet, currently, it is still blurry, but we give ourself one year to find a new brand for these activities.

The association would actually be one client of the cooperative. So if the governance of the association change, and we are not happy about each other anymore, the association would be free to find a new hoster (we hope not, but if we let the members decide, this is what can happen too).

This part is the part that currently pays the rent, obviously, and we hope to grow by next year. We have a goal to be 5 people by next year, but we’ll see :wink:
Actually, we don’t want to grow to infinity, we have a plan that if we reach 14 people, then we split the structure in 2.

We hope that the split is quite clear. If you have any feedback, please feel free to comment here.

β platform

We are working on a new version of the platform, and we want to release a β for 1st of February (5th anniversary and crowdfunding).
The idea is not crazy, it would be:

  • A cloud - with Nextcloud obviously :wink:
  • A collaborative note taking app - CodiMD
  • This very forum
  • Our chat
  • maybe a fediverse app, mastodon, peertube or funkewhale, we have to see what we can, and what you want :wink:

And everything behind a unified experience, with a single sign on, and some sort of dashboard. Details are still blury. We don’t plan to migrate the current shared Nextcloud, it would be too complexe. But we’ll keep it updated. We’ll also change the domain name, not so happy with .host .
And everything will be deployed in a Highly available fashion on our new infrastructure. And the platform is also easily scalable so we can scale to the infinity and beyond! We’ll probably be able to offer higher storage for lower price, but we’ll be able to do so, only if we have a bit of volume.
This β platform purpose is also to be a great demonstration for potential clients of the cooperative that would like to deploy such tool suit in their environment, under their domain name, with their logo, and their integration (ldap, already existing tools…)

α - libre.sh

We also want, for the 1st of February to release libre.sh in α version.
libre.sh is this very platform we are working since a year. This very forum is already running on top of it, and half of the infrastructure is already there. We are now working hard to release it free software. You can follow development here (not much yet, but it will come :slight_smile: ).

This idea of this platform is to be able to collaborate on how to operate hosted free software with High Availability, scale, backups, restore, self service for the end user, and with as little effort as possible for the admins. Since the beginning, it is our dream, and we see the building blocks getting together nicely.

If you are interested in that, follow this forum.

Hope you enjoyed reading, and we are really curious of your feedback! We know it is ambitious, but we think we can do it with you, and it is our duty to move free software faster!

PS: Here is the status of our infrastructure :wink:

  • old docker-compose infrastructure
    • 402 containers serving 104 applications on 4 servers.
  • test kubernetes cluster
    • 38 containers serving 3 applications on 3 servers
    • with a test ceph cluster
  • prod kubernetes cluster
    • 381 containers serving 152 applications on 6 servers
    • Backed by a ceph cluster of 30 TB hdd and 1 TB Nvme
2 « J'aime »

wow - thank you for this update! Really excited to read about your plans. They are ambitous, but sound very well thought through. :slight_smile:

2 « J'aime »

I am unsure what I should think about

If I pay 100€ a year for the IndiePass and something is not working, then I honestly think there should be some kind of official support. But I totally get your reasons behind your decision!


I am considering to switching to your « profesionnal service ». Will it look like your IndiePacks? Or is it best to wait for all the information until February when you do the crowdfunding?

And regarding

Do you already have some idea how this will be working out? 20 GB in the small IndiePack is a bit meagre, when you think about the fact that the IndiePass already will have 30GB.

Well if it is not working and it is an issue on our side we will fix it in a reasonable timing. We still plan to offer services that works ;). What we won’t do at all is user support, everything related to the use of the app should go on the forum (unless its confidential then contact us in private).

Actually in our model, indiehosters the association will be a client of the business entity, with some kind of contract that we need to define, and all the shared instances *.indie.host will be rock solid, its kind of our vitrine.
If you register for a dedicated nextcloud instance with the association, the service level won’t be the same though. We will be less reactive, if a server goes down, we will restart the services that took the professional support offer first. It won’t be high availability but of course we will always make sure it works.

Those professional services are targeting enterprises, it will be way too expensive for an individual. It is 400€/month for 1 app, 200€/month for extra apps + the cost of hosting (which is close to what we are doing with the indiepacks).

Indeed but the 30GB in the IndiePass are not only for Nextcloud but for all the apps that are in the pass.

With the last plans to split the activities, I think we will need to review our pricing and offers… again… :exploding_head:

We wont have customers anymore but members that register for a membership that gives you access to some services. That changes things. It will probably be something like a minimal (low) price to get membership and give more if you want. And extras, more storage for instance, will be charged.

1 « J'aime »

I think we will plan a « community call » very soon, so we can discuss about those things and get your feedbacks. We have been talking about this for a while with @pierreok but never did it.

As we mentioned, for now we keep control on the governance to kickstart the thing but those strategic decisions should not be only our call.

2 « J'aime »

Thanks for the clarification!

You mean if I get an IndiePack (my own dedicated instance) the service level will be « worse » compared to IndiePass (shared instance)? (Even though from a user’s perspective, one would be paying 2,5 to 6 times more than the most expensive IndiePass…?) Or the IndiePack is worse compared to the professional service, which is obviously quite clear.

Yes, that is right! Waaay to expensive! :joy: But now I do understand a bit more the difference between the cooperative and the association. First I thought that the IndiePacks - including professional service, high availability, etc - could be bought from the cooperative…

I’m going to do a bit of a brain dump here, hopefully it will clarify things and its also an open discussion.

Context:

We have 2 main segments in our users/customers.

  1. Formal and informal collectives, associations and individuals
  2. SME’s, coops, public institutions and big collectives

In both of those segments we have a big trend of people, mainly french, coming from communities contribution to the commons, les tiers-lieux (third place?)…

Problem:

We think that in the first segment people want to be part of a collective that mutualize resources, they contribute to the commons in reciprocity. While in the second they prefer to be a client that pays to get a service.

(this is not entirely true for all of the personas of those segments) They don’t expect the same level of service. For instance as an individual or a small association if my cloud is down for 2h it is acceptable. As a professional organization it is problematic. As an individual or association I can accept that my data could be lost in the « rare » case where I’m the unlucky one. I accept that it is my responsibility to mitigate this risk. As I professional I want to pay to mitigate this risk.

Of course the budget are also not the same.

To satisfy both those segments is a really difficult exercise.

Solution:

We took the the decision to split our not-for-profit from our commercial activities in an association. In the following model, the association is a client of the cooperative.

The association:

  • The association offers through membership a login to its shared services and/or access to dedicated instances.
  • The association is owned by its members/contributors who take all decisions
  • The association choose a hosting provider for its services

Teaser: we own https://libe.re domain, what do you think about https://cloud.libe.re ? (it is definitely french oriented)

The cooperative:

  • The cooperative offers dedicated instances with professional support.
  • The cooperative offers customized solutions and trainings
  • The cooperative manages and commercialized the products libre.sh (this will be another topic)
  • The cooperative is owned by its workers

The (new) problem:

For shared services it is clear, the membership is X€, you get a login to all shared instances. For individuals it should cover 95% of the cases. As an individual do I really need my own nextcloud?

Then it is actually quite tricky to define what would be the differentiator between the two offers in the case of dedicated instances.

In the case where I need a dedicated instance, I have 2 options:

  1. Cheap one, I get a membership for an indiepack from IndieHost/XXX association.
  2. Expensive one, I register for a professional hosting offer from IndieHost/Oxalis coop.

The risks:

We dump the market:

Our main worry is that the association’s becomes an alternative to the coop in a context where money flows. We don’t want to dump the market, we already did our part…

We consider it as an issue if I can choose between an indiepack from the association and a dedicated instance from the coop. If for 1% of my budget I can get the top of the top of the collaborative tools, this is not a good signal. We don’t want to contribute to this :slight_smile:

Cheap = Low quality

As it is the cooperative that is taking care of the technical infrastructure for the association, the level of service will be defacto quite good. We don’t want to spend time to make sure the level of service is worst… Either from the association or from the cooperative its the same infra.

Solution

So concretely ? … well its still an open discussion :upside_down_face:

Our last conclusions are:

No SSO:
The association won’t offer a Single Sign On solution on its dedicated instances. That would be a pro feature.

Standard Offer:
The pricing and offer for the IndiePacks in the association should be standard and quite straightforward.

We offer 2 types of applications (as regard resource consumption):

  • apps that need 2 workers - 15€/month
  • apps that need 1 worker - 30€/month

There is one single size, its designed for about 10 users (for discourse/peertube/… it would X views).

No disaster Recovery:
See Pierre’s topic about this: https://forum.indiehosters.net/t/data-conservancy-and-ecology/3362

tldr: Disaster recovery costs a lot of resources (x9 replications). Zero risk does not exist. The association will help people to manage data conservancy themselves.

For the low budget, contributors the common, I want to take care of my relatives… segment, it should cover 90% of the cases.

There is still a grey area, if you need more and/or pro feature. We still have cases where it is in-between the association and the cooperative with a low-medium budget, big expectations/needs but that « deserves » to be supported because it contributes to the commons.

I still don’t know if this kind of situation will be dealt in the cooperative or in the association or in the association then between the association and cooperative.

1 « J'aime »

Very interesting! Thanks for sharing your brain dump.

For me as non-French speaking person I would like https://cloud.lib.re more. The term libre is commonly used and would not hinder adoption outside of France. But lib.re is probably already taken :confused:

One of the main reasons to buy an IndiePack for me personally would be to be able to have more than one email adress. I would just use two accounts as a single user.


Bu apart from that - How are your plans evolving? I haven’t seen any crowdfunding in February. Is it still planned?