LWN: Comments on "Mesh networking with batman-adv" https://lwn.net/Articles/426947/ This is a special feed containing comments posted to the individual LWN article titled "Mesh networking with batman-adv". en-us Tue, 07 Oct 2025 09:50:51 +0000 Tue, 07 Oct 2025 09:50:51 +0000 https://www.rssboard.org/rss-specification lwn@lwn.net Mesh networking with batman-adv https://lwn.net/Articles/428801/ https://lwn.net/Articles/428801/ raalkml <div class="FormattedComment"> In the book "Rainbows End" by Vernor Vinge, the protagonists<br> _sometimes_ manage their mesh manually, adding and removing nodes<br> by explicit user request. I wonder if that may be a sensible approach<br> in mobile devices managed by their users, like the mobile phones.<br> </div> Fri, 18 Feb 2011 16:40:26 +0000 Mesh networking with batman-adv https://lwn.net/Articles/428542/ https://lwn.net/Articles/428542/ efexis <div class="FormattedComment"> Seems some learning from p2p networks like Gnutella could be used, where some nodes become "supernodes", used to maintain routing details for nodes further afield, whilst other nodes just maintain routing to nearby supernodes. When one node needs to talk to another, it can query its nearest supernode(s) for this information, and then cache it locally.<br> <p> Maybe :-)<br> <p> </div> Thu, 17 Feb 2011 15:52:55 +0000 Mesh networking with batman-adv https://lwn.net/Articles/427864/ https://lwn.net/Articles/427864/ showell <div class="FormattedComment"> The ever increasing mobile data speeds are causing an ever increasing need for more bandwidth by the Mobile Operators. This can only be met by moving up in frequency. We now see 2.6 GHz being deployed for LTE mobiles. What band do we use next? 3.5 GHz and 5 GHz are both possibilities but both these bands have very limited coverage. The answer at some point will be mesh networks.<br> <p> The focus on mesh in the Kernel is timely. One problem that needs to be solved is the time it takes for route discovery. If we do end up with "mobile mesh" then vehicles will be one of the most important nodes (large power source with decent antennas at a good height). The problem is that most of these nodes move at significant speed (when compared with the size of the mesh cell at 3 GHz and above).<br> <p> Low latency rebroadcast of messages or some smarter route discovery algorithm need to be priorities if mesh is to work in the next generation of mobile networks or beyond (5G or 6G).<br> </div> Mon, 14 Feb 2011 01:51:37 +0000 Mesh networking with batman-adv https://lwn.net/Articles/427435/ https://lwn.net/Articles/427435/ iabervon <div class="FormattedComment"> I think the desire to filter the content their device transfers is likely to be harmful to the intended use of the system; in a scenario where freedom of information is threatened, participants would want to carry state TV and sports in addition to regional news so that it's plausible (and even likely) that the reason for participating is actually to follow the latest football matches, not protest the government. That also means that the people who actually just want the latest sports aren't able to avoid spreading the region news, and everyone cares about football, while (in the situation where freedom of information is threatened and matters) only a significant minority of the population wants to protest the government.<br> </div> Thu, 10 Feb 2011 21:40:23 +0000 Mesh networking with batman-adv https://lwn.net/Articles/427256/ https://lwn.net/Articles/427256/ dw <a href="https://secure.wikimedia.org/wikipedia/en/wiki/Content-centric_networking">Content-centric networking</a> seems well suited for mesh networks, where instead of futilely attempting online communication across a network in continual flux, static information is opportunistically exchanged, with some ranking/reputability system used to decide which content to publish, and which to receive. Something closer to FidoNet than the Internet. <p> Another benefit is the possibility of not requiring battery-powered devices to be permanently awake in order to serve as router for other nodes in the mesh. It's easy to imagine a scheme where devices only wake for a specific window every few minutes to handshake with any interesting neighbour, synchronized to the device's current best idea of time (something that could even be derived from the mesh during initialization). <p> While it wouldn't allow you to watch every desired YouTube clip, in a scenario where freedom of information was threatened, such a network might be very effective for transmission of important information. One aspect that isn't clear would be how a user would configure their (space constrained) mobile device with a policy to filter the exchanged data, e.g. the ability to select regional news while excluding pornography, sports, or politically motivated content. <p> I think given software with the right UI, and some motivating application, technology already exists today in millions of pockets worldwide to make a network like this reality. Thu, 10 Feb 2011 05:31:36 +0000