Date: Fri, 15 Jul 94 04:30:02 PDT From: Advanced Amateur Radio Networking Group Errors-To: TCP-Group-Errors@UCSD.Edu Reply-To: TCP-Group@UCSD.Edu Precedence: Bulk Subject: TCP-Group Digest V94 #149 To: tcp-group-digest TCP-Group Digest Fri, 15 Jul 94 Volume 94 : Issue 149 Today's Topics: 800 mHz Trunked HT - To Person Interested Bridge vs. Repeater (2 msgs) Bridge vs. Repeater (fwd) DAMA v. Repeaters (4 msgs) PI/PI2 driver update Send Replies or notes for publication to: . Subscription requests to . Problems you can't solve otherwise to brian@ucsd.edu. Archives of past issues of the TCP-Group Digest are available (by FTP only) from UCSD.Edu in directory "mailarchives". We trust that readers are intelligent enough to realize that all text herein consists of personal comments and does not represent the official policies or positions of any party. Your mileage may vary. So there. ---------------------------------------------------------------------- Date: Thu, 14 Jul 94 11:35:39 CDT From: route66@ddl.chi.il.us (System Administrator) Subject: 800 mHz Trunked HT - To Person Interested To: tcp-group@UCSD.EDU I know this is off the topic..bt.. To the persn interested in my 800 mHz Trunked HT, and wanted them. I hav lost your address so I'm unable to FedEx them to you. If you will email me your address at route66@ddl.chi.il.us Thanks, N9TOL De Greg ------------------------------ Date: Thu, 14 Jul 1994 09:12:06 -0400 From: goldstein@carafe.tay2.dec.com Subject: Bridge vs. Repeater To: tcp-group@UCSD.EDU >Well, isn't there a bit more than that? Even if two nearby users don't care about routing, a router can keep two local users from sucking up bandwidth... I'll share this with the group... Routers and repeaters solve different problems. When you maintain your car, do you change the oil or the check the tires? Waddayamean both? What kind of religious war are we having if we have two different correct answers? A repeater -- operating in the physical layer -- is VITAL to solving the hidden transmitter problem. I personally think DAMA's impractical for most ham use, for various reasons including the crappy radios we glue on to, but a repeater WILL get rid of HTS. When you have Hidden Transmitter Syndrome, the actual channel capacity degrades to about 18-20% of bit rate, since it's effectively Aloha technology. When you have a repeater and everyone hears everyone else when they're NOT transmitting, then you have CSMA and the channel capacity increases dramatically, MORE than doubling, though the actual efficiency depends on various timing factors. When you have a repeater and everyone else is ALSO full duplex, then you have CSMA/CD and the channel capacity increases even more. I've seen Ethernets run over 90% capacity, even though the theory says that you can't run over say 60% or so -- the theory assumes WORST CASE with everyone at maximum distance in a huge star. So a repeater uses two frequencies but will get MORE bps/Hz than the crappy Aloha digis we have now. It's insane that we don't use them on 2 meters more; haven't the old yakkers died off yet :-)? The old first-come first-serve policy is enough to warrant giving the band over to commercial users, almost, since it guarantees the obsoete old stuff will forever prevent more efficient newer users. A router works at a higher layer, between users who do NOT have any physical layer connectivity (or at least that's it's main functiom). That's just not the same function as a repeater. Now if the router were atop a mountain, it might choose not to pass along stuff that didn't need tocross the mountain. But what does that solve? The presence of one user's packets crossing the mountain will cause bits to be present on both sides, and local users will get clobbered anyway. The mountaintop should probably have a repeater for the local users AND a router to go to OTHER mountaintops. fred k1io ------------------------------ Date: Fri, 15 Jul 1994 12:24:20 +1000 From: makinc@hhcs.gov.au (Carl Makin) Subject: Bridge vs. Repeater To: tcp-group@ucsd.edu At 10:57 AM 13/7/94 -0500, David Rush wrote: > Because traffic between two stations on the left side of the mountain would > generate useless traffic (thru the regen) on the right side of the mountain. > (Sucking up bandwidth.) A router could decide which packets need to cross > over the mountain, but otherwise stay quiet. Whether you want to route or repeate depends on the traffic flow on the network and the topography of the surrounding terrain. Canberra (Australia) is quite hilly and the BBSs are at the north and south ends of the city. The biggest problem is hidden transmitters. The terrain means even in the same "valley" many stations can not hear everyone else in the same "valley" and the majority of the traffic on frequency is going to or from the next "valley" anyway. In this case a bit regen repeater solves a number of problems and as the traffic is mainly crossing between "valleys" then we don't lose bandwidth. We also don't have enough local traffic to justify the expense of isolating the "valleys" and linking them via routers (yet). We've chosen to make the entire ACT one "LAN" (for 4800 baud traffic anyway). > And when we say "bit regenerator", it's functionally equivalent to an > Ethernet repeater, aren't we? Yes. -- Carl Makin (VK1KCM) "Speaking for myself only!" makinc@hhcs.gov.au 'Work +61 6 289 8443' Canberra, Australia 'The best book on programming for the layman is "Alice in Wonderland"; but that's because it's the best book on anything for the layman.' ------------------------------ Date: Thu, 14 Jul 1994 09:05:19 -0500 (CDT) From: Kurt Freiberger Subject: Bridge vs. Repeater (fwd) To: tcp-group@UCSD.EDU (TCP-Group Mailing List) rush@dns.sprintcorp.com said: > Well, isn't there a bit more than that? Even if two nearby users don't > care about routing, a router can keep two local users from sucking up > bandwidth... I'll share this with the group... First of all, one needs to determine the topology of what one is trying to do. You have to define "local". In most cases, "local" means LAN. The audience is considered peers in one group, with one shared media. It'd be nice if we had one channel for one path, but physics ain't with us. A router inter- connects subnets. A bridge interconnects LANs as peers within subnets. > I said: > >> Yeah, presumably cross- or multi-banding would probably make the most > >> sense (easiest). All of the stations in a "mutually hearable area" > >> would share one freq on one band. The stations in a different > >> "mutually hearable area" (on the other side of the mountaintop > >> router) would share a different freq on a different band. > > Then Gerry said: > >So, why not use a wide area machine on 2 in-band frequencies, using the > >regen? It's a "repeater" to the FM guys, simple to maintain, simple to > >build, simple to install. Find a good {tower/building top/mountain top} > >and put it up to cover as much of the are as possible. If you THEN have > >an area that's not adequately covered, put up another regen in that area > >on the reverse channel, or route over to that remote area. > > Because traffic between two stations on the left side of the mountain would > generate useless traffic (thru the regen) on the right side of the mountain. > (Sucking up bandwidth.) A router could decide which packets need to cross > over the mountain, but otherwise stay quiet. It depends on whether you need the other side of the mountain to be a different subnet. The decision to use a single regenerator (repeater) on the top vs two regenerators with an interconnect is one of economics. Obviously, two radio sites with interconnects is more expensive. If you have two regenerators, then you can decide whether to route or bridge according to the LAN topology. To subnet or not to subnet, that is the question? > Assuming that nobody even cares about routing in this case, what I'm > really proposing is a bridge (to use Ethernet terminology) between > the two sides of the mountain, rather than bit regenerator. > And when we say "bit regenerator", it's functionally equivalent to an > Ethernet repeater, aren't we? You have it exactly. I've been mulling over the design of a bridge for AX.25, and how that could work in our environment. There are some DEFINITE gotchas, especially with regard to bridging loops. I didn't really want to deal with spanning-tree at all!!!! 73/Kurt -- # Kurt Freiberger, WB5BBW Dept. of Computer Science, TAMU # # Internet: kurt@cs.tamu.edu | "Even MY hypocrisy has its limits." # # AuralNet: 409/847-8607 | # # AMPRNet: wb5bbw@wb5bbw.ampr.org | - "Doc" Holliday, Tombstone # # Disclaimer: Not EVEN an official document of Texas A&M University # ------------------------------ Date: Thu, 14 Jul 1994 11:50:09 -0500 (CDT) From: ssampson@sabea-oc.af.mil (Steve Sampson) Subject: DAMA v. Repeaters To: tcp-group@ucsd.edu goldstein@carafe.tay2.dec.com writes: > I personally think DAMA's impractical for most ham use, for various reasons > including the crappy radios we glue on to, but a repeater WILL get rid > of HTS. Well I think a crappy radio will not perform either DAMA or Repeater duty, so I guess I don't understand your logic. Why DAMA won't succeed in most Western societies is because we are raised to always seek power in our sphere. That means every Ham will want to be the Master Node, and not submit to being a Slave. Here in Oklahoma City we have more Nodes than users because that's how people are. If you ask them if they want to create a network the answer is: NO! They just want to beacon their Callsign and own a frequency. Oops, I've drifted from technology to behavior again. . . sorry. -- Steve ------------------------------ Date: Thu, 14 Jul 1994 14:20:11 -0400 From: goldstein@carafe.tay2.dec.com Subject: DAMA v. Repeaters To: tcp-group@ucsd.edu > I think a crappy radio will not perform either DAMA or Repeater duty, so I guess I don't understand your logic. Why DAMA won't succeed in most Western societies is because we are raised to always seek power in our sphere. That means every Ham will want to be the Master Node, and not submit to being >a Slave. Nope, that's not why I think DAMA won't cut it. First off the average radio doesn't HAVE to do repeater duty; a CSMA (no /CD) arrangement is probably the happy medium; tha'ts ONE repeater and ORDINARY HDX radios. DAMA itself is based on polling, right? So in DAMA, you need to make FAST polls to potential users. Long Txdelay times make it fail, and that's the problem. If there are 50 users on the freq. and only 3 are active, you avhe to waste time poling 47. For a wireline tolkien ring that's no issue; they take microseconds. But voice-oriented ham radios take many milliseconds to be polled. Way way too many. ALso, the protocols we use are not master/slave; or are they very happy about waiting long times for a round-robin poll to take place. Super-fast polling would fix that but I just don't see it. Other than that, Mrs. Lincoln, how did you like the play? fred k1io ------------------------------ Date: Thu, 14 Jul 1994 13:55:45 -0500 (CDT) From: ssampson@sabea-oc.af.mil (Steve Sampson) Subject: DAMA v. Repeaters To: tcp-group@ucsd.edu > DAMA itself is based on polling, right? Yes. > So in DAMA, you need to make FAST polls to potential users. Long Txdelay > times make it fail, and that's the problem. Well it would be a problem compared to NO txdelay true. > If there are 50 users on the freq. and only 3 are active, you have to waste > time poling 47. Well, as I understand it - When you are polled and respond with IDLE, the Node will begin backing off on your poll timing. So the 3 active will get the channel while the 47 IDLE users will get more and more backoff. Then if one more user switches from IDLE to ACTIVE they get back to high priority again. I guess this is akin to Prioritized Round Robin in Task scheduling. -- Steve ------------------------------ Date: Thu, 14 Jul 94 14:22 EDT From: nelson@crynwr.com (Russell Nelson) Subject: DAMA v. Repeaters To: tcp-group@ucsd.edu From: goldstein@carafe.tay2.dec.com Date: Thu, 14 Jul 1994 14:20:11 -0400 DAMA itself is based on polling, right? So in DAMA, you need to make FAST polls to potential users. Long Txdelay times make it fail, and that's the problem. If there are 50 users on the freq. and only 3 are active, you avhe to waste time poling 47. For a wireline tolkien ring that's no issue; they take microseconds. But voice-oriented ham radios take many milliseconds to be polled. Way way too many. But if only three are active, why waste time polling 47? Why not make a poll priority list, doing an exponential (up to a point) fallback in priority. Also, why should the hub be doing any transmitting? It should just assign slot times for the users, assigning more slots to more active users. Or is DAMA some sub-optimal scheme that someone is pushing as a universal solution? I'm afraid I've got my "Answer for every question regardless of my ignorance" hat on. -russ http://www.crynwr.com/crynwr/nelson.html Crynwr Software | Crynwr Software sells packet driver support | ask4 PGP key 11 Grant St. | +1 315 268 1925 (9201 FAX) | What art thou doing about it? Potsdam, NY 13676 | LPF member - ask me about the harm software patents do. ------------------------------ Date: Fri, 15 Jul 1994 00:00:22 -0400 (EDT) From: dp@pluto.va3dp (Dave Perry va3dp) Subject: PI/PI2 driver update To: tcp-group@ucsd.edu A driver update is available which improves performance for the B port on PI and PI2 cards. Depending on the host machine, it should now be capable of 9600 baud (on 386 class machines). Also included is a bugfix that should eliminate rx overruns. This update is for the NOS source code. The Linux driver update will follow. It is available for ftp from hydra.carleton.ca as: pub/hamradio/packet/tcpip/pi2/pi.c Dave va3dp ------------------------------ End of TCP-Group Digest V94 #149 ******************************