Bitcoin Forum
August 30, 2026, 08:46:33 AM *
News: Latest Bitcoin Core release: 31.1 [Torrent]
 
   Home   Help Search Login Register More  
Pages: [1]
  Print  
Author Topic: [ANN][EDX] EdgeX Network — Edge-Native Proof-of-Work, Mined on Your Own CPU  (Read 80 times)
edgexNetwork (OP)
Newbie
*
Offline

Activity: 1
Merit: 0


View Profile
August 29, 2026, 03:14:35 PM
 #1

[ANN][EDX] EdgeX Network — Edge-Native Proof-of-Work, Mined on Your Own CPU

Website: https://www.edgexnetwork.org/
Explorer: https://export.edgexnetwork.org/
Pool: https://pool.edgexnetwork.org/
GitHub: https://github.com/edgexNetwork/EdgeX-Network

=====================================================================

EdgeX Network is an edge-native proof-of-work blockchain that puts
hashing power back where it belongs — your own hardware.

Mainstream PoW drifted far from the "one CPU, one vote" ideal:
ASICs captured nearly all global hashrate, full-node clients grew
heavier, and mining turned into a game played by industrial
datacenters. EdgeX flips the equation by pairing the ASIC-resistant
RandomX algorithm with the Bun/TypeScript async runtime.

With EdgeX, every idle core counts. Home desktops, NAS boxes,
Raspberry Pis, and other small edge devices can mine and run a full
node — and every node doubles as an edge service provider (DePIN).

=====================================================================

■ KEY FEATURES

• RandomX consensus — CPU-optimized, hard memory-dependency rules
  out ASICs and large GPU rigs.
• Bun Native FFI — the C++ RandomX kernel is bound to a lightweight
  TypeScript node with near-zero overhead. Full node boots in
  milliseconds.
• Micro-node engine — embedded SQLite storage, no external database,
  one command to start mining and validating.
• P2P gossip network over TCP + WebSocket, with NAT traversal.
• Built-in JSON-RPC and Stratum mining protocol support.
• EdgeX Wallet TUI — terminal wallet with BIP39 mnemonic, encrypted
  vault, QR codes, i18n, and daemon mode.
• DePIN-ready — nodes can expose decentralized RPC and caching
  services to the public.

■ TOKENOMICS

• Ticker: EDX (smallest unit: Photon, 1 EDX = 10^8 Photons)
• Hard cap: 2,100,000,000 EDX
• 100% fair launch — zero premine, every coin comes from CPU work.
• Block time: 15 seconds; LWMA difficulty retarget (240-block window)
• Three-phase emission:
  - Bootstrap (days 1–90): 400 EDX/block
  - Smooth decay (day 91 – year 10): continuous decay formula,
    no halving cliffs
  - Tail emission (after year 10): 1.5 EDX/block + protocol-level
    fee burning

■ HOW TO GET STARTED

Mining & running a full node is one command away.

Windows (PowerShell):
    irm https://install.edgexnetwork.org | iex

macOS / Linux:
    curl -fsSL https://install.edgexnetwork.org | bash

Launch the node/wallet:
    dexcoin

■ LINKS

Website    : https://www.edgexnetwork.org/
Explorer   : https://export.edgexnetwork.org/
Pool       : https://pool.edgexnetwork.org/
GitHub     : https://github.com/edgexNetwork/EdgeX-Network

EdgeX — PoW belongs at the edge. No farm required.
zygzag
Newbie
*
Offline

Activity: 20
Merit: 0


View Profile
August 29, 2026, 04:43:29 PM
 #2

discord ? telegram ?
morlok007
Newbie
*
Offline

Activity: 3
Merit: 0


View Profile
Today at 06:10:39 AM
Last edit: Today at 06:28:44 AM by morlok007
 #3

Heads up EDX miners — I found a serious bug in the node, please read before you commit rigs

Hey all 👋

Like some of you I got curious about EdgeX (EDX) and dug into the v1.0.0 code
(https://github.com/edgexNetwork/EdgeX-Network) before pointing hardware at it.
Unfortunately I hit a serious consensus bug and I'd feel bad not sharing it — better
you hear it now than after burning time and electricity. Purely technical, no drama,
no claims about anyone's intentions.

What's wrong
In block validation (packages/core/src/state.ts, applyBlock), the node
validates every transaction against the state from before the block, and only
mutates the UTXO set afterwards (the code even says so: "Mutations happen only after
every transaction in the candidate is valid"). It de-duplicates transaction ids, and
inputs within a single tx — but it never tracks coins already spent by
earlier transactions in the same block. So you can put two different,
properly-signed transactions that spend the same coin into one block and both
pass. The mutation step then deletes the input once but creates both sets of
outputs. In plain terms: you can mint EDX out of thin air.

Why it matters
It breaks the 2.1B hard cap — balances and total supply can't be trusted, and anyone
who mines a block can inflate at will. Fixing it is a hard fork + fresh genesis, since
blocks accepted under the bug become invalid.

I didn't just eyeball it — here's the proof
Ran against the project's own validation code, offline and in memory (nothing done to
mainnet), using the same state.clone() the real accept path uses. I also checked the
whole path — acceptBlock -> addBlock -> PoW check -> state.clone().applyBlock — and
there is no double-spend guard anywhere outside applyBlock.

One 400 EDX coin, two different signed txs spending it:
Code:
tx1 alone        -> ACCEPTED (recipient 399 EDX)   <- each tx is individually valid
tx2 alone        -> ACCEPTED (recipient 398 EDX)
tx1 + tx2 block  -> ACCEPTED (recipient 797 EDX from a single 400 EDX coin)
Each transaction is valid on its own, so this isn't junk slipping through — it's two
legit spends of the same coin, both honored. 797 out of 400, in one block. Repeat at will.

The good news: it's a small fix
Track spent inputs across the whole block during validation:
Code:
const spentInBlock = new Set<string>();
// ...in the per-input loop:
const key = utxoKey(input.txid, input.index);
if (spentInBlock.has(key)) throw new UtxoValidationError(`double spend within block ${key}`);
spentInBlock.add(key);
I applied exactly this and re-ran: single txs still accepted, the two-in-one-block
attack is now rejected. No false positives.

One more thing — the "run your own node" story doesn't hold up
I checked how you'd actually join the network, and it's a mess:
  • The release ships six prebuilt binaries and they're ALL wallets
(dexcoin-wallet-{linux,darwin,win}-{x64,arm64}). There is no node binary. I
inspected the Linux build: it contains no P2P, no RandomX, no mining/validation
code — it's a wallet that just talks to a local node RPC. So the advertised
"curl | bash then run one command to mine and validate" actually installs a
wallet only, not the full node it promises.
  • The open-source node (from source) ships with zero seed nodes: EDX_SEEDS is
empty by default, no DNS seeds, no hardcoded peers.
  • No peer/seed address is published anywhere public — not the README, the ANN,
the website or the pool page.
[/list]
Net result: from every public artifact, you cannot actually sync a node or join the
P2P network — you'd need a peer address handed to you out-of-band. For a project whose
whole pitch is "run a full node on any edge device," that's a big gap. Not a
vulnerability like the double-spend above, but it means right now you can't
independently verify the chain even if you wanted to.

A couple of smaller things
The P2P transaction handler isn't wrapped in try/catch (the block handler is), so a
malformed gossiped tx could crash a node. And the node RPC has no auth — fine on
localhost, risky if exposed. Credit where due: the RandomX binding looked clean and
there's genuinely no premine (genesis issues nothing, coinbase amount is validated),
so the core isn't sloppy everywhere — this one bug is just a big one.

My honest take
I'd hold off putting rigs or money on EDX until this is fixed and the chain relaunches.
Not trying to FUD anyone — I just don't want people burning power on a chain whose
supply can be forged. Devs: happy to hand over the full reproducible PoC. Mine safe out there o/
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.19 | SMF © 2006-2009, Simple Machines Valid XHTML 1.0! Valid CSS!