These practical problems make the project less feasible:
1. Making end applications implement your protocol (eg Facebook) makes it way harder to scale, simply because the enshittified big tech apps are unlikely to care. Now the apps will have to maintain a separate service and not get paid to do it.
2. Having deliberate physical couriers travel across borders are a massive risk - a hop to hop mesh network where connectivity can be easily established across all users (every user is a courier when they connect to unlimited mobile data or wifi) will make the network a lot more available
Even after reading the technical explanation I don't understand how such a network would be useful or who the couriers are and how they actually deliver the cargo... Or why that's desirable over just caching the messages until the internet is available again. And what about reliability? How long does cargo take to get to its destination? How often does it actually make it there?
Hiya! Author of Awala here. We're revamping the website right now so it's great to get this kind of feedback! I'll answer your questions in the meantime.
You raise good questions. I originally tried to keep it simple and wrote the documentation with my (prospective) partners in mind. They already work in this space, so they know the problem very well, and we'd usually have a few high-level conversations before getting into that kind of details, so I didn't want to bore them with things they already knew. However, things are changing now and we're opening things up to the public in the coming days, so these are things I need to document better.
> how such a network would be useful
We're trying to establish connectivity in regions where the national/local government, or an adversary in a conflict zone (e.g. Gaza, Tigray), deliberately cuts off the population from the Internet.
I'm talking about places where satellite Internet services, like Starlink, won't work:
- Conflict zones where the enemy controls the sky, as they could "triangulate" the location of the terminal. This is why Starlink can be used in most of Ukraine, but is too risky to use in other conflict zones.
- Regions where the government won't give a licence to the operator. For example, Starlink in parts of Sudan: https://www.theguardian.com/global-development/article/2024/...
> who the couriers are and how they actually deliver the cargo
A courier is an individual or a group of people who volunteer to transport the data physically, between the region without access to the Internet and a location with access to the Internet. They can charge people if they want to and people agree, but Awala itself doesn't handle anything to do with that, so it'd be more of a verbal agreement amongst them.
That location with access to the Internet can be a place within the disconnected region that the government is intentionally keeping online (e.g. hospitals, international hotels), or a place/subregion whose ISP is taking too long to take offline. Worse case scenario, it could be:
- A place near the border with another region/country, where they can use the 4G/5G cell towers from that place. For example: https://www.economist.com/middle-east-and-africa/2019/01/10/...
- Another region that the couriers can travel to. For example: https://www.aljazeera.com/features/2020/1/13/how-internet-ex...
> Or why that's desirable over just caching the messages until the internet is available again.
Because in many cases you never know when the Internet will be restored. And during that time, you have diaspora communities absolutely horrified after not hearing from close relatives for weeks or even months. Or, in the case of many North Korean escapees, potentially ever.
> And what about reliability? How long does cargo take to get to its destination? How often does it actually make it there?
Every situation will be different. For example:
- In non-conflict zones where the government doesn't have the capability to triangulate unlicensed Starlink devices, it'd take whatever it takes for the courier to drive/cycle/walk to that device. People have smuggled these things even in Iran, where the government does have some capability to detect them: https://iranwire.com/en/technology/133773-iranians-defy-inte...
- In places like India (the world's capital of Internet blackouts), where blackouts are regional, couriers could catch a train to another region as shown in the link above.
Awala is built to withstand delays of up to 6 months.
Hi there. I’m from India (the world’s capital of Internet blackouts). A major reason for Internet blackouts here is because in many areas, there’s a deep seated animosity between different communities due to historical and ideological differences. Internet blackouts are done close to any sensitive event, to prevent malicious actors from spreading misinformation/rumours and provoking riots. Have you considered the ethical implications of your service in such cases? Are you willing to take the moral responsibility for the damage to life and property that could be caused using your service?
They simplified the explanation a bit too much in some places and lost details in the process.
I think this is the kind of connectivity that works in places with no other connection, like a village on top of a mountain, an area with long term power outage, places where internet is censored, etc. Some of these cases come with a risk to the courier. The couriers could be for example the regular buses connecting that place to a hub with internet connection. The equipment on the bus "caches" your requests or replies, takes them to the connection hub where it picks up whatever comes back.
Hello! Author of Awala here.
> They simplified the explanation a bit too much in some places and lost details in the process.
Please let me know if there's anything else that we could improve, besides the questions raised to the comment by Vampiero that you're replying to.
> I think this is the kind of connectivity that works in places with no other connection, like a village on top of a mountain, an area with long term power outage, places where internet is censored, etc
Indeed. It could work in those cases, but my priority are regions disconnected to the Internet due to government orders or a foreign adversary.
This is more of an anti-censorship thing. Most useful in areas where somebody turns off the Internet for political reasons.
A really reliable low-bandwidth emergency network would be more useful. Something that forwards short text messages via phone to phone WiFi until they reach some place with more connectivity would be helpful in emergencies.
The challenge with out-of-band RF networks in certain situations is that they can be triangualted by a hostile power. In places like Ukraine, that can trigger indiscrimante artillery or rocket fire onto civilians, for example. In China that can get people disapeared. Sometimes sneakernet is safer and more secure.
> This is more of an anti-censorship thing. Most useful in areas where somebody turns off the Internet for political reasons.
Yup, that's exactly the only use case I'm targeting (I'm the author of Awala)
> A really reliable low-bandwidth emergency network would be more useful. Something that forwards short text messages via phone to phone WiFi until they reach some place with more connectivity would be helpful in emergencies.
That's also useful, but I'm not sure about either option being more useful than the other. These are very different problems within the realm of offline comms.
It would be more useful to build a bluetooth mesh network (like what airtags already uses) to carry small pieces of information across long distances without internet access instead of having physical couriers that need to deliberately stop and have people connect to them to store data and potentially charge a fee
It should be ubiquitous and something that just works when someone passes by
This sounds a lot like Scuttlebut
Hello! Author of Awala here. Not sure about that. We're trying to solve very different problems.
Scuttlebut and SSB-powered apps like Manyverse focus on social media and decentralisation, which is more about "broadcasting" and gossip protocols, where the offline comms is a happy side-effect of their decentralisation goals.
Awala is a platform to make software communicate offline securely. Centralised services like Twitter[1], and decentralised ones like Letro[2], or even hybrids, are all possible.[3]
I'm not on a quest to decentralise the Internet. I'm on a quest to connect people to their loved ones despite repressive regimes and wars.
[1] https://github.com/AwalaApp/poc#relaynets-proof-of-concept (Awala used to be called Relaynet)
[3] https://awala.network/tech-overview#service-decentralisation
>Behind the scenes, your Awala app connects to an Internet gateway, which is a server that acts as a bridge between your device and the rest of Awala on the Internet.
It's literally just a proxy.
When the Internet is available, that specific component in your quote (the Internet gateway), does _generally_ behave like a proxy.
The interesting thing is what happens when the Internet is unavailable. Check this out: https://awala.network/tech-overview#sneakernet
And this video in particular: https://awala.network/assets/diagrams/centralised-couriers.m...
- Gus (author of Awala)
Well, hello there Richard Hendricks! ;)
Author of Awala here! You have no idea how much my friends tease me with that! :)
how would this work in jurastictions that are likely to jail you for transporting information across borders like say the great firewall of china etc.
I would expect that just having the app installed would get you some quiet time in a small room.
> how would this work in jurastictions that are likely to jail you for transporting information across borders
Great question! This is, by far, my biggest worry. Especially in North Korea if we ever launch there. This is actually going to be the topic of the next episode in the Inside Awala podcast: https://awala.alitu.com/
Long story short, there are measures that we've taken, and measures we will take.
Measures that we've taken include using another layer of E2E encryption for the data transported by couriers and minimising the metadata to such an extent that you couldn't see who's the sender/recipient, even if the authorities took the courier's device to a lab. This way, couriers couldn't be coerced into giving something they don't have.
Measures we will take include concealing the app. For example, it could look and behave as a calculator when you open it (even its icon and name could convey that), and you'd only see the actual app once you enter a particular math expression.
Also, the technology is getting independent security audits periodically. They're currently wrapping up the latest one, and the report should be available in the coming weeks.
As for China, that's a totally different problem: They do have access to the Internet, but it's heavily censored. However, we're tackling that problem too now, as we're adding a resilient VPN function to Awala, whose prototype was already tested in China: https://github.com/relaycorp/fanqiang-poc
The reason for this VPN functionality inside Awala is that, at the end of the day, every single region that's susceptible to Internet blackouts is also subject to Internet censorship when the Internet is available. So people shouldn't have to install multiple apps for the various types of censorship they will be subjected to at different points in time.
Have you considered leveraging steganography in the sneakernet transport layer? Maybe something like https://steganography.live/info
China is very special case of totalitarianism with low entropy, but most other totalitarianism cases have low entropy only in capital City, and extremely corrupted outside capital, especially on borders.
This is don't mean that you could run across border on station wagon filled with prohibited content, but somebody could hide something like microSD in clothes.
Exists paradox - people usually thinking, jails have very strong and serious security system so inside there should be sterile environment without anything prohibited, but in reality usually there exist black market of prohibited things.
But for that, just what you communicate would be checked and why communicate…