Feasability of Ethereum on top of the pinoccio?

StephanTual

If you don't know the wonderful pinoccio, you're missing out: https://pinocc.io/

I'm wondering as to the feasibility of implementing Ethereum on top of it. This would give instant mesh-network capabilities - reminiscent of the bitcoincard (http://bitcoincard.org/)

Jasper

You mean running Ethereum ontop? It is an ATmega256RFR2, 16MHz MCU certainly not going to run a full node.. Probably not a lightweight one either. But they probably are good for other things, for instance if they can securely talk to something with a bit more computing power, memory and storage.

Of course something like that is probably able to hold the private keys, so that they're not actually on your computer, and are hard to steal. Even better if you could view the transaction to check that it is actually what your client is claiming it is.

erictj

Super cool to consider Pinoccio <-> Ethereum in various projects!

I did want to mention that we're working hard to move our networking stack over to Telehash (http://telehash.org), which makes much more sense in the context that you're discussing, since it is itself a completely decentralized protocol. In short, it'd let any device (even the small Pinoccio) connect to any other device that also supports Telehash--regardless of location or type of network connectivity. And all this is done with no centralized server requirement.

We expect to have the Lead Scout <-> API using Telehash in a month or so, and Scout <-> Scout using Telehash later this year.

Happy to discuss how this might work with Ethereum if you'd like.

-Eric @ Pinoccio

plummerfernandez

Hi eric, I'd like to take up your offer to discuss how this might work with Ethereum, your product is fascinating. thanks for your help.

PhonikG

So RasPi/Pinoccio integration is one option here, in order to run the eth node(RasPi) and still utilize the mesh(pinoccio) - http://support.pinocc.io/hc/en-us/articles/201464300-Frequently-Asked-Questions#faq-rasppi

Any thoughts/updates on this topic @erictj ?

Bitcoinzie

@PhonikG From what i can tell, if you can get an eth node running on the pi then you can connect your pi directly to a pinocc.

guttertec

I do not know if you guys heard about theHybridGroup (Cylon.js etc.) or Meshblue (use to be called Skynet). I could imagine to get those guys fascinated on a port towards the Blockchain. If you want, have a look at my research… http://research.axelquack.de/iot/2014/09/18/swarmRobotics.html

PhonikG

erictj said:

Super cool to consider Pinoccio <-> Ethereum in various projects!

I did want to mention that we're working hard to move our networking stack over to Telehash (http://telehash.org), which makes much more sense in the context that you're discussing, since it is itself a completely decentralized protocol. In short, it'd let any device (even the small Pinoccio) connect to any other device that also supports Telehash--regardless of location or type of network connectivity. And all this is done with no centralized server requirement.

We expect to have the Lead Scout <-> API using Telehash in a month or so, and Scout <-> Scout using Telehash later this year.

Happy to discuss how this might work with Ethereum if you'd like.

-Eric @ Pinoccio

Eric, is it true that pinocc.io is a dead project?!? A shame if it is... I got the impression there was alot of interest. I myself would purchase a kit if it was ramped back up with proper support!

tym4t

All it needs is to be able to talk over Ethereum's P2P (swarm), a working port of libsecp256k1 for you to do some interesting things with it. You just have a contract running that can send it messages.

linagee

Use something more standard like an esp8266 instead. (I loved pinoccio when it launched, but esp8266 has taken over.) Way faster than a pinoccio as well. Instead of 32kb of ram or whatever, there is around a meg of ram and 160Mhz in a $3 module. Arduino support as well through "esp8266/arduino" project on github.
Not sure if this still would need to be just an RPC client or if it could dig deeper and be some sort of "ultra-light node"... (At least it should have enough processing power to make sure nobody is "pulling the wool over it's eyes" and verify a single transaction or something.)

tym4t

Even the proposed SPV will require some processing power and P2P access over the network so unless this is powered by a battery pack it won't make it. It has to do bloom filter verification, look at transactions that it is interested in, traversing the state trie to verify.

A more clever approach would be to have several of these things running in their own trusted mesh scheme each of them having their own cryptographic identity and distributing the verification steps between them in some random fashion. You have a given number of node look at bloom filters, a given number of nodes traversing the trie, a number of nodes requesting blocks from full nodes, then collect the results.