What you describe can be done but it would be a step back, not forward from Packet networking as we know it. - This assumes that you are talking about a network that covers a significant area and has a good number of users in it. Your idea would be most advantageous with a small number of users within a small area. For something like that, you would be best off utilizing old WIFI equipment, off of the ham bands. - And we all know how far the development and utilization of ham radio WIFI went... When it was finally realized that PART97 was not going to be repealed, allowing internet content to be irresponsibly transported over the ham bands, all of the interest in ham radio WIFI suddenly dried up and blew away. - R.I.P. hinternet.
There has been very little development of Packet radio networking for some years and absolutely none of it in the USA, a legacy of TAPR's decline and its protocol wars against packet networking that followed. The most advanced and sophisticated packet networking system at this time is still FlexNet, which was developed quite a while back in Europe, where they are fortunate enough ( as Packet networkers ) to have no equivalent of the fallen TAPR organization. You could approach your idea on VHF by having all users operate FlexNet node software and only utilize the 'backbone' of such a network. FlexNet is the only system developed for ham radio that will automatically recognize new nodes, route around dead or slow spots as they appear, and re-route to utilize resources that speed up again or reappear, all without user intervention. This would be a step backward though, in both performance and reliability from simply utilizing FlexNet as intended, with user accces ports at a slower speed than a seperate set of backbone links. To illustrate the basic principle: If user access is "speed 1" and backbone links are "speed 8", then sixteen halfdup users can simultaneously utilize the network for point to point links without being aware of any network-related slowdown. This is called 'transparency', as the users do not detect any slowdown related to other traffic on the network. As more users utilize the network after that, transparency suffers. In the system you describe, as more than two users utilize the network it becomes progressively slower right away, until it finally becomes unusable. This is why I said it would only work for a small number of users, all utilizing the highest speed equipment available. Economics would dictate that you to settle for the short-range WIFI equipment for this, so you would pretty well be limited to a small number of users all located within a very small area. Far better is the utilization of FlexNet as put forward by its developers. - Node to node links would be the only utilization of the high-speed, long range full-duplex coinnections which would be far too expensibve and complicated for each user to reasonably expected to operate. Since each node serves a fair number of hams, the expense of its high-speed, long range ( UHF ) links can be shared and thus not offer an insurmountable burden. - This allows users to utilize slower, less expensive and complicated halfdup equipment at home to access the nodes, and for much of the time they will enjoy network transparency. The last I heard, the European FlexNet network was moving away from 9.6kb access toward 19.2 halfdup, with 78.x ( 78.6? ) and faster fulldup backbone links. If you want a fast, reliable, transparent Packet network, this is the way to go. - It is proven and it works well for point to point links over dozens of nodes to destinations seperated by hundreds of miles or more. If your main interest is in transporting internet content over the ham bands though, then forget it. - The 'hinternet' concept is dead. Common sense and reason prevailed, which is why you do not hear much about the ARRL HSMM group anymore. 73 DE Charles, N5PVL ----- Original Message ----- From: Miroslav Skoric To: [email protected] ; [email protected] ; [email protected] Sent: Tuesday, January 06, 2009 12:50 AM Subject: [digitalradio] AX.25 modem as an infrastructure? Hi all, A description of One Laptop Per Child project (ACM Queue, Nov/Dec 2007) suggested they had implemented a kind of a mesh networking which means a laptop is also the infrastructure for 'relaying the Internet' ("If one laptop in a village is connected to the Internet, they all are.") What makes me wonder is if our end-user packet stations can behave in a similar manner, which means to implement a 'silent node' within AX.25 modems and TNC's - so whenever a new station appears on a frequency it could serve as a network 'node' for its surroundings (with, say, up to 10% of node traffic at Level 3/4 or so). That way we might not depend so much on fully-equipped nodes and, in the same time, to implement a system of 'dynamic node networking' where our packet signals could travel via best routes possible at the particular time. Is something like that a viable idea? (I know of a digipeater on/off option in most TNC's but what I suggest seems to me as a more complex approach.) Any existing solution that already works that way? Regards, Misko YT7MPB
