[ANN] GRAHAMBELL - Capped PoW Mining to 1 Hash/Sec/Node | No Parallel Mining

[zee47]

Yes you did. Your interactive demo validates it. Trying to figure out how!

[Hurairah Shamsi]

Updated the post body.

[Hurairah Shamsi]

Added an explainer video on YouTube that shows how the Local Client works and how Proof of Witness and Witness Chains work together to reject PoW mining attempts that exceed 1 hash per second per node:

https://youtu.be/i5gzzqFXXUk

[Hurairah Shamsi]

Added another video link: Visual representation of PoW mining during audio/video calls.

https://youtu.be/gPz7d3lGz1k

[Hirarchy]

Just watched the demo. Pretty impressive to see the project prioritizing decentralization. This initiative has massive potential; with 1 hash per sec per node, the low barrier to entry could enable widespread participation. With everyone running full nodes, it avoids centralization concerns of blockchain.

[yoshikiazuma]

ive been watching this and im interested. can't wait to try this out

[Hurairah Shamsi]

We are now live on Product Hunt.
Would appreciate support and feedback from the community.

Link: https://www.producthunt.com/products/grahambell?utm_source=other&utm_medium=social

[Hurairah Shamsi]

Q/A

I got asked the following question:

Interesting approach.
If each node is capped at 1 hash/sec, how do you prevent Sybil scaling (running many low-cost nodes) from reintroducing parallel advantage at the network level?
Is identity or resource binding required, and if so, what enforces it?

It’s a fair question, so sharing the answer here as well.

Registration itself is a Proof of Work process.

Before a node can mine transaction blocks, it must first mine a registration block containing unique metadata. The hash of that block becomes the node’s ID. Without a valid ID, the node cannot propose transactions and peers (registered) will not connect to it. Registration is therefore enforced by consensus.

Registration blocks follow the same global throughput constraints as the rest of the network. They are bounded by block time, not by local hardware.

Similar to how you cannot mine 1 million Bitcoin blocks instantly just by adding machines, you cannot generate unlimited node IDs in parallel beyond the network’s registration throughput.

If registration averages ~1 minute per block (example here), then producing 1 million valid IDs would take ~1 million minutes (~694 days) on average, assuming zero competition.

Sybil scaling therefore becomes a time bottleneck rather than a hardware bottleneck.

Witness chains then restrict parallel connectivity both during and after registration, but that is a separate mechanism from the registration throughput constraint.