Hey There,

I'm a systems administrator (by career), and interested in networking/redundancy personally as well. I could never find alternative methods of linking up remote computers outside of using the internet, this is when I found out about packet radio, and amatuer radio in general, I'm (slowly) working towards my license..hopefully this summer. At first only networking over shortwave radio interested me, but it's really starting to grow on me.

Any how I read the whole HSMM-MESH site, and documents, and it REALLY intrigued me, especially the redundancy it provides, allowing people to still be connected regardless of internet, as well as having an excuse to setup linux mail/voip/file servers at home ;). Two questions I did have were: What kind of bandwidth is theoretically possible between nodes in the mesh, I know this is a question that can't be answered definitively as it depends on many variables, but I'm curious if it's 100kbit, 1mbit etc..

The other question I had was what are the typical distances you have seen between nodes? Another "it depends" answer I know, but would love to hear the distances anyone has gotten.

Thanks!

P.S. Anyone in the Ontario, Canada area?


EDI: I found the slideshow and read that, looks like 40KM range (best case) with dual 19dbi antennas. and speeds were listed too :)

I am also curious if anyone has made measured latency tests.

Although I just found out about the HSMM-MESH project I have been working on a similar network in west Kentucky for a couple of years.  We are not running MESHing software yet but I am interested in doing so.  I teach in the Telecommunications Systems Management program at Murray State University and I have build the beginnings of a regional network as a way of providing lab experience for my students in my Intro to Wireless Technology course.  I also teach Intro to Telephone Technology and will follow with great interest the VoIP experiments here.

We have been taking a bit different approach here.  We're using 5.7Ghz 802.11n for backbone links using directional antennas but keeping the patterns pretty broad by using backfire or panel antennas (17 to 20dBi at 90 or 120 degrees).  The goal is to multiply connect the nodes and let either OSPF, some form of meshing or both handle connectivity changes.  The idea is for these links to be able to take a fair amount of shaking and still be reasonably aligned.  These links are operated as part 15 devices so we don't have restrictions on encryption, etc.  We control access to these tightly.  We can make this work very reliably at fairly high transfer rates (32Mbs) over paths up to 25 miles though we're trying to keep most to 20 miles for fade margin purposes.  These connect Access Nodes on 2.4Ghz 802.11g/n that use omnis (12 - 15dBi) or multiple panel antennas (15 to 17dBi) to provide user access.  Still working on the ranges for the access nodes but reliable omni range should be in the 3-5 mile range using omnis and up to 15 miles using grids or backfire antennas.

I received funding for some links to study "Network Survivability of Distributed Decentralized Networks". We are looking at network survivability defined as being able to transfer "useful" amounts of data for emergency comms rather than restoring full network capability.  Basically, we're going to measure latency, throughput and connectivity on a network like I described in the earlier post and try to determine what the minimums are for useful communication.  I'll the posts here updated if anyone is interested.