Meshkit | Hacker News Reader

Meshkit

opengarden.com

112 pointsjdowner33 comments

This is very unfortunate piece of technology. And I'm not saying this ironically.

This is the type of technology that will benefit the most number of people if open sourced, but despite what these guys call themselves ("Open"garden), it's not open at all.

I'm guessing it's because they've already raised near $13MM https://www.crunchbase.com/organization/open-garden and they can't afford to just let go of the tech (I'm guessing they're probably hoping at this point to get acquired by some large company for their tech, but not even sure if they have a patent-worthy tech since I'm guessing they use iOS and Android API underneath) If they didn't raise this much money they could have made the decision to just open it up if they really think this will change the world.

As much as I hate ICOs, I would like to see this type of tech being open sourced and the creators trying to make money with open protocols rather than getting stuck in the old mindset of trying to make money by petty licensing in the age of open source.

Also, note that there's a better alternative if you're trying to implement near field P2P AND you're OK with using someone else's proprietary tech https://developers.google.com/nearby/connections/overview

I looked into this field a while ago and found that there's nothing out there that's completely open. For a genre of technology that claims to liberate communication, it's super ironic how it's the opposite.

We have tackled a protocol to solve ad-hoc mesh network data transmission in P2P environments.

It is MIT/ZLIB/Apache2 Open Source licensed: https://github.com/amark/gun

I explain how the P2P logic works in this tech talk: https://www.youtube.com/watch?v=5fCPRY-9hkc&t=3s

We also have end-to-end encryption now working: https://github.com/amark/gun/wiki/auth

What we're missing, is we need adapters for Bluetooth, etc. Andre Staltz (of CyclesJS fame) has already added React Native bindings ( https://github.com/staltz/gun-asyncstorage ), which I'm hoping some of those tools (like Phonegap, Nativescript) have bridges for Bluetooth?

That was a fantastic talk; thank you for the link.

For anyone else confused (it might just have been me) but the URL is correct, title of the talk is "Nordic.js 2017 • Mark Nadal - The Design and Evolution of Event-Driven Databases" or "How to build a global scale, P2P, mesh network?"

That is an honor to hear, thank you for the feedback. :) :)

Was there anything I didn't address, or didn't have time to go into more detail on, that I can answer for you?

I have to be honest, for an idiot like me it was absolutely perfect.

Here's what we're trying to achieve, why we're trying to achieve it and some nice blocks that work towards doing that (with the idea that you make them modular enough and can bypass them).

I can only imagine a real world implementation ends up having all the edge cases in the world, but you gave me a great insight into what some of them are likely to be :)

There is also Gotenna [1] and numerous projects mentioned on Wikipedia mesh networking pages [2, 3]

This would be a perfect application for "Shadow Dance" - a long shot project to create futuristic festivals inside the shadow path of total solar eclipses including multimillion person human chains (see the technical proposal - particuarly page 7 - [4]). The idea came too late for the August eclipse but our group is planning and hoping for the Mexico US Canada eclipse in 2024 - (the website and the proposal is in the process of being rewritten accordngly)

1 https://en.wikipedia.org/wiki/GoTenna 2 https://en.wikipedia.org/wiki/Mesh_networking 3 https://en.wikipedia.org/wiki/Wireless_mesh_network 4 https://drive.google.com/file/d/0B5eQ1ipl4xCDQVF6ajVNalg3Wjg...

Indeed, my understanding is that neither iOS or Android provide reasonable access to adhoc or 802.11s modes on their wifi chipset. I gather MeshKit is deploying its mesh topology through tunnels, somewhat abstracted from whichever phy.

The ideal would be a company with funding that releases an open source product. None of the open source technologies that are community driven have taken off, to my awareness.

You have to realize that all these BigCo driven open source projects are really exploitation.

I would rather see an ICO model--where the community funds the protocol, like FileCoin or Ethereum--succeed than Google or Facebook taking advantage of the naive developers.

You have to realize that all these BigCo driven open source projects are really exploitation.

In what way? Other than Android and maybe Chromium, what open-source products are exploitative?

Man, spearheading a FOSS project that could do this would be a dream, but I struggle having enough time to tinker as it is.

Nice idea, but despite the "Open" in the name of the company, this is all based on a proprietary protocol. The internet has done pretty well so far by being based on open protocols and standards. I don't think it's a good idea to change that.

Yikes, their case study showing a map with pinpoints for DNC protestors in Philadelphia is an alarming illustration of the capacity for these peer-to-peer comms apps to be abused. I wonder if the FireChat participants at the time were aware of the location tracking. https://www.opengarden.com/case-studies.html

Does not seem like a free and open standard so my tab was closed in seconds.

Sounds a bit like http://www.servalproject.org

Or Briar: https://briarproject.org/

Or Telehash: http://telehash.org/

Why do none of these projects manage to publish a usable embeddable library (+ docs)? That could be so much more impactful than all the minimal demo-chat apps built on top of the mesh network tech...

Development in a vacuum, everyone wants to cash in, disagreements on protocols or languages or politics or people.

I’m a big ham radio nut, and although the premise of the hobby is the advancement of radio technology, everyone’s old, has their own way of doing things, and has reinvented the wheel many times over. Logging programs, DX clusters, remote radio interfaces, APRS programs and apps, satellite trackers, you name it, there’s a dozen versions of everything.

I imagine this is common in technical hobbies.

Does this allow iOS and Android to talk to each other over the mesh without reaching into the cloud? It says iOS and Android, but it doesn't seem to say -between- iOS and Android.

is there any (sdr) usb radio that could be connected to android phone via usb-on-the-go to reach devices within 1+ mile on unlicensed bands? to be used by this or other open source mesh implementation, like http://servalproject.org/

Something like this would work great in an urban environment if enough apps used it. For non-urban environments the wifi range becomes a limiting factor. I'm hoping that LTE Direct becomes a thing, since it has a much higher range.

Kind of a separate idea, but if the hardware ever became cheap enough, it'd be great if the charger that shipped with these "LTE Direct" phones also included an LTE chip so that it could participate in the mesh. It would be great for disasters and also for overloaded network

That way you'd have a decent chance in a suburb to be able to have a no-data-plan-necessary mesh network for devices of preteens or other minor uses where paying for a data plan is too expensive. For example, a data connection for your entertainment system in your car.

In order to connect such a mesh to the Internet without having to worry about being the origin of someone else's traffic, the base stations would connect over wifi to their home access point. All traffic would be relayed over the Internet to the user's own base station, which would then send it to the destination. That way the mesh doesn't have to have a lot of traffic but someone can connect to the mesh without becoming the point of origin for potentially malicious traffic. It also would reduce the load on the mesh since traffic would only travel one or two hops.

Something similar is accomplished with http://thaliproject.org/

Very interesting - we've been looking at something very similar (though less extensive) for bridging between BLE end devices and our cloud. It'd be interesting to see what it'd take to get ble devices integrated and on the mesh (battery requirements/mem requirements).

Right now we're just forwarding beacon packets to our database servers - but this seems to provide a back/forth that'd be very useful (command/control & ota updates)

Difficult to say without more detailed docs

I think this is a really cool idea.

Is there a way for one to limit the amount of data that could go through your device so that your data caps don't suffer from sending too much of somebody else's data?

Edit: I see the "User control" feature, but it's not too specific about what exactly publishing and subscribing means for you vs strangers on the same mesh.

I contacted Open Garden, interested to develop a product based on their purported architecture that would be a useful in-the-field extension of my established SaaS product.

The website indicates they see themselves as a solution vendor for specific verticals. My application isn’t in the use cases on their list.

I did not hear back.

So much for MeshKit.

I am more excited about this project https://github.com/AntonTheDev/ExtendaBLE

Disclaimer: Created by one of my co-workers.

I like this idea. I keep hoping that a solution with peer to peer networking can help us overcome the problems we have with ISPs owning and controlling all of our access to the internet.

I want to see this working. I want to try it out myself. Sadly, I'm not part of a multi-million customer company, so I don't know if I am the demographic they are looking for.

If you like the concept, check out http://ipfs.io, great team & concept

Not the same thing at all. IPFS doesn't do any mesh networking it handles file distribution, storage and retrieval but doesn't do any of it's own networking so it relies on some existing connectivity from what I can see.

I am pretty sure this also uses some existing connectivity tech such as Apple's multipeer connectivity https://developer.apple.com/documentation/multipeerconnectiv...

At least IPFS guys open sourced everything for anyone to use. The "Open"garden people had the guts to name themselves "open" despite it being opposite of open.

That's not an existing link though until two devices using it come close enough to communicate and the connection is established. If that were to count any device with an ethernet port and an IP stack would be linked because "all" you need to do is run a cable. Of course both devices have to have compatible protocols to communicate but that doesn't mean they have a link established until the connection is made.

IPFS is really neat and I'm glad it's open sourced but it's not a mesh networking protocol. I'd love to see a combination of IPFS and a robust mesh networking scheme. It'd be a pretty great way to bring internet like functionality where there's no infrastructure without having to setup anything other than having people bring their devices.

Reply on news.ycombinator.com