Beam a good support for a "plug-host"?

As an internet user I get bored by the lacks of the Internet/Web telematic (centralized) architecture and would need some true local capacitation (“multimatic”) - and smart standardized (or not: AI) dialoguing with the related hosts. This could be provided by a local virtual host I could “plug” into my browser (cf. OPES RFCs). I identify that it should be OS independent, capability oriented, and able to support what elixir provides. I do not yet program in Erlang/elixr and I hear about erlang2. Question1. would beam be a candidate to install, support, and run plug-hosts? If yes I will start learning and developing in elixir (same for erlang?). Question2. Would erlang2 modifiy some of the capacities of Elixir to support such a project? Question3. Such a user oriented project is to be an open-source project. I am considering an open Foundation to support it: do you think it could be of interest among the Elixir community.
PS. the concept is (personal, automation, security, etc. oriented) so the key is robustness, reliability, and flexibility. Less speed.
Thank you for your comments.

In my opinion you could edit your topic and make each question to appear in a separated paragraph to make it easier to read, thus improving the possibility to get more people interested in the discussion.

1 Like

I’d absolutely be behind an effort like this but you’ll have to give us a bit more details on your ideas (which do sound good).

Is this about cars ? I have no clue what this is about. I must be lacking some keyword because my google searches shows nothing very interesting.

The RFC’s are over 15 years old. https://tools.ietf.org/html/rfc3835

Who provides these services? What is the use-case?

I see mention of proxies https://tools.ietf.org/html/rfc3752

What sort of protocols are these based on?

Do they require long running processes to maintain state?

1 Like

In this case, I will spend more time on the concept in an Erlang/Elixir context. But, my Point of View is currently strictly internet architecture-oriented. I just know how I would develop it under QNX, and from my participation in the OPES WG. I just want to understand if a local “Erlix” virtual machine could support what I have the need for. (A few years ago I tried to set-up a local meet-up on elixir but did not raise real interest).

1 Like

Thx for the advice. My idea was just to raise the matter, first question, being embeded in a context to help understand it was a project, not only a question in the air. The reply of dimitarvp tells me that the question is not entirely of the scope.
What I am going to do is to set-up a working site on the concept. So I may structure ideas and received comments/suggestions. Being understood I am entirely new irt. erlang/elixir: I hope from that to get/(may be introduce ?) some new ideas.

1 Like

Actually, this is a very old thing. It is genuinely architectural to diginumeric networks. Including neural networks.

This was my area for the international public operators packet switch system (IPSS) that was strategically obsoleted when the IETF internet took over in 1986. It is more or less related to the OSI Presentation layer n°6 which is missing in the TCP/IP model. This embarasses us a lot (multilinguism, smart networking, security, capabilities, etc… and network negentropy are missing). Web agents are a part of a response: RFC 1958 provides the architectural solution, which states “The network’s job is to transmit datagrams as efficiently and flexibly as possible. Everything else should be done at the fringes.”.

The problem was to understand how to introduce the related “extended network services” at the “fringe” (what it exactly is) in the TCP/IP architecture. The great idea was to introduce them on the flow (at the edges) and not at a node (what would have changed too many things).

However, you are right: this has not taken off yet as we still all are TCP/IP “confined”. This is why the first thing is to do it within the TCP/IP architecture. I feel HTTP and JMAP are now stabilized enough for us, “interlinked-users” (there is no “end” in an internet system) to invest our “BTW” (“beyond the wall”, i.e. citizen) R&D on the matter. Because we need it in order to live together out of GAFGov control.

I therefore started http://rfc.wiki to concentrate and translate the RFCs I feel concerned.

It is making the network smart through network flow negentropy. Post-shannonian network. You receive more than it was sent to you, and they receive more than what you typed - but as you wanted it. In a more secure manner : this is named digital capacitation/augmentation capability.

I was not saying you were out of scope at all… What I was asking you to do was something around this:

As an internet user I get bored by the lacks of the Internet/Web telematic (centralized) architecture and would need some true local capacitation (“multimatic”) - and smart standardized (or not: AI) dialoguing with the related hosts.

This could be provided by a local virtual host I could “plug” into my browser (cf. OPES RFCs). I identify that it should be OS independent, capability oriented, and able to support what elixir provides. I do not yet program in Erlang/elixr and I hear about erlang2.

Question1. would beam be a candidate to install, support, and run plug-hosts? If yes I will start learning and developing in elixir (same for erlang?).

Question2. Would erlang2 modifiy some of the capacities of Elixir to support such a project?

Question3. Such a user oriented project is to be an open-source project. I am considering an open Foundation to support it: do you think it could be of interest among the Elixir community.

PS. the concept is (personal, automation, security, etc. oriented) so the key is robustness, reliability, and flexibility. Less speed.

Thank you for your comments.

See how it reads much better, thus easier to understand and spot the different questions you are asking :wink:

Deep thx. This is exactly what I am to paste in the website to prospect the project. It will be in mediwiki, but I long to know how to build it under elixir. Thx!

We are starting to move out of TCP:

QUIC will be used for the new standard HTTP3:

1 Like

We crossed paths before and I still haven’t replied to your very interesting email from several months ago. Sorry! Life hasn’t been kind to me this year and things just started finally getting better like two weeks ago for the first time.

Know that I am very interested in your ideas – although you should probably work on translating them to a simpler human language because not many of us have such an excellent academic background as yourself, which can be a hurdle in the communication.

I still would like to see several examples of what your ideas would bring about in the world. We chatted a bit and I very much liked your idea about the Wiki and a broader inter-connected knowledge base working together.

2 Likes

Sorry, yes I forgot that GAFGov are surrendering to Google, so “TCP/IP” may lead to misunderstanding. Let say the “IETF” architecture (refitted UDP) as per its original father: Louis Pouzin. The problem is not there, it is in the missing layer six presentation. The beauty of the OPES is that the concept works on the “flow”, meaning on datagrams themselves. As long as Vint Cerf does not want to get rid of them, working on the flow is OK. However, and this is why I introduce the “plug-host” concept, the last spot IETF will not change is the “fringe” on the user side of the plug. cf. RFC 1958. Because the flow can be unreadable (as on Tymnet) or encrypted.

Hi!
I know but things have to chain together. The S7 (scientific summit of the G7, sciences academies) has identified two layers of citizen science and research : participative and “beyond the walls” of laboratories (http://frhm.fr/index.php/Citizen_science_in_the_Internet_era). I am focussing on the idea that the BTW layer (absolute top level R&D, because not hampered by budgets and policies) needs a smart-pack to live, work, and interknow at best cost. I feel that an “erlix” beam based architecture could provide such a system. But I need to deep into all the aspects of this - to be also in agreement with the networked way of working/thinking of the “citizen technosophy”. So we can trust it ourselves ! My targets are what you describe: but I need to validate them in making them criticized by high/deep level experts, and for that to challenge them. :slight_smile:

I completely agree here… I am struggling to follow the conversation.

2 Likes

I also fully agree. But please understand that I come from the debate at the IETF, I left because they have decided to only be the technical support of ICANN/GAFGovs - while I am on the user side. Their investment strategy is in the Internet BUG (i.e. being standardised by Beeing Unilaterally Global). This is such that we can only have our way outside of their areas of control. This means on our side of the fringe (RFC 1958) and on the unprotected flow (OPES). My idea is that we can do it with a single development; as a smart OPES plughost, because they will never affect the datagrams between our socket and our plug. But we need to be entirely consistent for the plughost to work on any uncrypted flow for it to be trusted and used - also as a way to control the smart-what GAFGovs or hackers may want to feed into our systems.

I will work on online images/maps this week !

2 Likes

Could you give an example of what that would be used for. I guess it would help people like me that do not follow :slight_smile:

Yes. Please give one or two days. I will have a web site explaining it.

1 Like

sent a mail to the list yesterday, but it does not get through. Saying that the plughosting issue tied to OPES was a wide/exciting matter calling for more work than I expected to keep it “kiss”. Additional time requested. Thx.