Has anyone experimented with alternative tunneling solutions? My concern or observation is that devices appear to have a max of 10 client/host connections and depending on the node size this can be reduced farther. While I understand a tunnel is not meant for permanent, long term solution I personally plan on maintaining two separate meshes (one local radio and one based on tunneling.)
The concept in an emergency weather locally or assisting others in an emergency I could patch one to the other. If my local area is an emergency and needs more resources and I still have some sort of internet I can alow my local node to patch though the tunnel nodes. As a reverse If my local mesh is running just fine and another location is in need of mesh resources but has internet I can provide mesh resources this way. The point is I believe the tunnel is an important tool to utilize but not dependent on.
This said maintaining two (or more?) mesh networks is ideal for my situation. Has anyone tried something like softether or something similar to tunnel to another network? I just began looking at it and it looks like (unconfirmed) you can disable the encryption as well as other beneficial features such as 802.3ad channel bonding (redundancy and increased throughput?) This would alleviate the 10 connection stress on the node itself with something as simple as a raspberry pi.
I guess my question would be would this have potential or would it be a waste of time/resources, and would it benefit the community as a result?
The concept in an emergency weather locally or assisting others in an emergency I could patch one to the other. If my local area is an emergency and needs more resources and I still have some sort of internet I can alow my local node to patch though the tunnel nodes. As a reverse If my local mesh is running just fine and another location is in need of mesh resources but has internet I can provide mesh resources this way. The point is I believe the tunnel is an important tool to utilize but not dependent on.
This said maintaining two (or more?) mesh networks is ideal for my situation. Has anyone tried something like softether or something similar to tunnel to another network? I just began looking at it and it looks like (unconfirmed) you can disable the encryption as well as other beneficial features such as 802.3ad channel bonding (redundancy and increased throughput?) This would alleviate the 10 connection stress on the node itself with something as simple as a raspberry pi.
I guess my question would be would this have potential or would it be a waste of time/resources, and would it benefit the community as a result?
N2MH uses OLSR (and other software bits, presumably) on a Raspberry Pi to create a beefy tunnel server. I'm starting to investigate that on this coast, as our three tunnel servers in the county are about tapped out. I've asked him for info on it (because I don't want to recreate the wheel) and will post whatever I hear.
Orv W6BI
I would caution that this was not the intent of creating the tunnel feature. Tunnels were designed to be temporary links to mesh islands until such time as RF links could be brought into the area. From a technical perspective, we have no idea what will happen to OLSR when networks get real large. At a minimum they will:
I certainly get that configuring a large, global network would be cool... but that's where it ends for me. I discourage you from building these on the AREDN platform.
Andre, K6AH
•tend to congest the network with OLSR updates
Hopefully by keeping the Internet constellation of nodes separate from the radio archipelago of nodes I am hoping this will not be an issue.
• not work when the Internet fails (like in most disaster situations)
That is one of the issues I really hope to overcome with an alternative tunnel. A system that has multiple fail safes could potentially mitigate issues. I have been toying with cellular connections, bonding connections (ie cable and dsl) for fail over and would like to explore satellite internet (I know i have read bad reports) but as a last resort how will it perform?
• give those Ham groups planning to support disasters a false sense of security that their networks will actually support the needs of their served agencies
Again what I am envisioning is not a main function of what aredn does, but rather supplement it. I would run it on a separate node than my radio links. 2 separate systems 2 separate purposes. In theory any nodes connected to mine via radio would never cross or communicate with the internet connected mesh unless proven stable. Plus a lot of us have no known nodes around us. For me (for example) the closest node is over 100 miles over uneven terrain. While I will be placing a node in hopes to gather traction in the area, I would have zero nodes to interact with.
Just like a lot of aspects of ham radio we have to experiment, police ourselves, and be ready. We follow guidelines help the community become stronger. I have absolutely no intention of being a part of a negative impact of this project Just really excited.
Another use for tunnels (which we've had to use unfortunately) is to route around RF link outages to maintain local network continuity until repairs can be made.
Orv W6BI