V-PACK v1

[RUBENQUANTUM2]

Hello Cisco DevNet Community,

My name is Ruben. I am an independent researcher from Ukraine.

I am sharing the first public version of a project called V-PACK and I am looking for technical specialists who are willing to examine it critically.

This is not an advertisement and I am not selling a product.

I am bringing this work to several technical communities because different specialists see different problems. I am interested in criticism from people working with networking, distributed systems, storage, AI infrastructure, system architecture and data movement.

V-PACK v1 is the first public implementation of an alternative approach to representing structured state.

It should not be understood simply as file compression.

The central question is different:

Does every logical state used by a computational system need to exist physically in a fully materialized form?

A conventional system may store, copy and transport explicit state together with other state that can potentially be reconstructed from known relationships.

V-PACK explores a different representation in which information can be separated into:

independent physical state,
relationships and dependencies,
and state that can be deterministically materialized when required.

The purpose of version 1 is deliberately limited.

I wanted the basic representation principle to be public, testable and falsifiable before building a much larger architecture on top of it.

I am therefore not asking anyone to believe that V-PACK is better.

I am asking specialists to try to break it.

Look for incorrect accounting.
Look for hidden costs.
Look for reconstruction bottlenecks.
Look for workloads where conventional representation clearly wins.
Look for architectural assumptions that do not survive real systems.

A demonstrated failure is useful to me.

I am writing the repository name as plain text because the purpose of this post is technical review, not promotion.

Version 2 is intended to go considerably further.

V-PACK is being integrated into a larger stateful computing architecture that I am developing around GEN9 / R-Core and a cognitive system currently called Yoshi.

The objective is not simply another LLM wrapper or another chatbot.

I am working toward a persistent, adaptive, multimodal cognitive architecture with its own state, memory, relationships, dynamic interactions and long-term development.

AGI is a useful shorthand for part of that direction, but I do not mean “an AGI chatbot”.

The intended architecture is broader: a persistent computational entity whose learned state and behaviour are not fundamentally dependent on a conventional Transformer weight-matrix execution model.

This is a development target, not a claim that such a system has already been achieved.

V-PACK v1 is simply one component that can already be inspected independently today.

Why am I posting this in Cisco DevNet?

Because Cisco engineers work in exactly the areas where the physical movement, representation and synchronization of information become expensive:

networks,
data centres,
distributed systems,
telemetry,
edge infrastructure,
control planes,
storage traffic,
and AI infrastructure.

If there is a fundamental problem in the V-PACK approach, specialists working with these systems should be particularly good at finding it.

That is exactly what I want.

Please challenge the architecture.

If an assumption is wrong, tell me which one and why.

If a particular workload destroys the advantage, show me the workload.

If the representation creates a new bottleneck somewhere else, point to it.

I am much more interested in a technically justified objection than in praise.

One final clarification:

This request for technical review is not exclusive to Cisco.

I am discussing the work in more than one technology community because I want independent criticism from specialists with different backgrounds.

I am not using this post to advertise those communities and I will not fill it with links to them.

The technical question is larger than any individual platform.

Regards,

Ruben
Independent researcher
Ukraine

Language note:
The original ideas, technical content and position expressed in this post are mine.

AI was used to translate and adapt my Ukrainian text into more natural technical English.

[r.heitmann]

Hi Ruben,

just to help me understand - we're not talking about this repository https://github.com/Alathra/VPack/releases right?

Please give a link to your github-repository.

Thank you in advance!

[RUBENQUANTUM2]

Hi,
No, that is a completely different project with a similar name. 🙂
The repository you linked is Alathra/VPack, a Velocity/Minecraft resource-pack plugin.
My V-PACK project is here:
https://github.com/etoyruben-wq/V-PACK⁠�
This is the repository discussed in my post and contained in the attached ZIP.
Thank you for pointing out the name collision — I can see how that could cause confusion.

[Jesus Illescas]

I'm inclining to mark this post as spam (will let other to comment as well), however you were clear that's not your intention. I feel the post and your project is a bit off-topic for what we do in DevNet.

I reviewed your github as well, and is not straightforward what your project does or intend to do, I consider is a file compression algorithm but your readme says otherwise, so is confusing. 

I consider is an interesting topic, given compression algorithms are an important piece of infrastructure that don't get a lof of attention like with the xz scandal, but is a bit off topic from this community in general.

[RUBENQUANTUM2]

Hi Jesus,
Thank you for taking the time to read the post and look at the repository.
First: if you genuinely believe this should be marked as spam, you are free to do so. I am not asking Cisco for special treatment or an exception to the rules.
But I would like to clarify my intent.
I did not come here to advertise a product. I came here as a developer looking for engineers who are willing to examine an unconventional architecture, criticize it, and tell me where it fails.
I live and work in Ukraine. Missiles and Shahed drones are part of everyday reality here. I do not mention that for sympathy. I mention it because under these conditions I have very little interest in spending years worrying about whether an unconventional technical question is perfectly comfortable for everyone in a warm office.
I build, test, publish, and look for people capable of attacking the engineering rather than simply saying “this looks unusual, therefore block it.”
There is also one important misunderstanding in your reply:
V-PACK is not intended to be a file-compression algorithm.
If it looks like one at first glance, then I have clearly not explained the architecture well enough.
The public V-PACK release is only the first visible component of a larger stateful/nonlinear computing architecture. Its purpose is not simply to make files smaller. It explores a different way of representing physical, derived, and relational state.
So if you did not understand what it is intended to do, I would genuinely invite you to look again and challenge the actual architecture.
Tell me:
where the state model is wrong;
where reconstruction becomes more expensive than explicit representation;
where the accounting fails;
where the architecture becomes impractical for real infrastructure.
That is the discussion I am looking for.
If DevNet is genuinely the wrong Cisco community for this subject, I would appreciate it if you could point me to the Cisco community where engineers working on data representation, distributed state, infrastructure, AI systems, or network architecture would consider it relevant.
I am presenting this work in multiple technical communities because I am actively searching for specialists, not an audience.
And I am equally straightforward about collaboration: if a company or technical community has no interest in examining the work, I simply treat that route as closed when I later consider technical cooperation, licensing, or access to related patented work. There are many engineering communities in the world.
Criticism is welcome. Technical disagreement is welcome. Cooperation is welcome.
I only ask that the project be judged for what it actually is before it is classified as something it is not.
Regards,
Ruben