The Baud Rate Race...

I Love Packet radio I love the internet but to me they are 2 differant worlds. I remember when the backbones were up, it was nothing to get up in the morning and chat with a friend in Chicago over coffee. What I saw was a bunch of 1200 baud nodes (at peoples Houses)going up that had no purpose. I had a 1200 baud node that tied into the KGM3 backbone node it worked great and my node was at 9600 feet. I am currently working to get may node back up and hope to get someone in calif to put a bacbone link into Az again. Packet is radio part of being a ham internet is for everyone. Packet Forever..Its great fun.

 

i can see the viability of that type of network, but... as large as it would span, i could not really forsee using anything but IP4... which would be a good place for ip4 once ip6 takes over the "Wired World" there will be gateways of course... but a segregate network would be good for things such as a thread i just posted to, about Anonymous Posters... oh my! this would be a GREAT venture in my book... who's to say it can't be redistributed locally over 802.11a, b OR c even? i do alot of work in the recent times with that stuff, and find it's very fast, but limited in it's robustity... even 900mhz will work on fractional watts, with no line of sight...
I'd be down with a medium or advanced speed digital network for several purposes i'd find useful, and local distribution would even be better as an appendage, to thin out the load on the slower backbone! perhaps a "caching proxy" at each jump point would be in order as well, gee we're looking at like 300 bux worth of junk here, with the 802.11b and a 486 to buffer requests.... the power requrements would take about a 12 volt battery every couple days, a solar and wind generator would be able to, oh, nevermind about that... that will be later...

Yes... very good ideas from all, as... Local communities would have ALOT more to transfer than someone across the country... and as such, multiple layers, nodes, and distributions would be in order, however, THE BACKBONE is a must!!! without it, it would be hard to get around u know? to cover a 40 mile radius at even as low as 900mhz would take 50 times the investment of a regular plain jane packet network, in my thinking... don't quote me, i have no knowledge or figures to back me up... just propogation awareness... 802.11b is FLIMSY at BEST...

here there's an internet company, starting up out in the country on 802.11b, and expanding to all the rural communities in my area in the NEAR future... they're not doing it right, but the company that buys them, will be in for a treat.. but still, with my friends coroberation, we've determined that their range is Overestimated by FAR... so they got a couple 200Foot towers, hey, anybody hate when their DBS system goes out? hehe this is 100 times worse! i would opt for 900mhz and below as a basis for this type of network... and there's some pretty reasonable DSL speeds available at that...
Multi-linking, Muxing, load-ballanced data transfer... all these techniques could spawn a whole new industry for data networking as we know it!! i'm excited about the prospect of this, and as soon as i got a 200 foot tower next to the radio shack i will soon be furnishing, you can COUNT ME IN!! i plan to build this tower to withstand at least 100 antennas and arrays of all sorts... so no ham band will be left out, of course, i'll have to be commercially sound on the project as that magnitude of building will require capital to maintain.

i have more, but it's past my bed time... all in favor? all in not favor.. who cares... let's just do it! see what happens

 

Although I tend to agree with N5PVL's personal assesment of Greg and Phil, they didn't cause the demise of the Packet Networks.  The packet networks died because they were slow, boring, and next to useless.  

Packet BBSes were pretty much joke grade in their heyday and are just totally laughable today.   Keyboard QSOs are much more interesting on HF digital and don't depend on infrastructure.   DX Clusters are at least useful locally but depend on the Internet (nobody needs two day old spots).

APRS is a (the?) bright spot in Packet Networks but seems to have come along to late save them on a large scale.

On the otherhand the Europeans have found an actual useful role for packet, Internet access.    Their networks are productive and thus are supported by a critical mass of hams.

Arguing about protocols is pointless unless one has something productive to do with the protocol.   Who cares if you can talk to no one more efficiently.

I think that this is the legitimate basis for the arguement for TCP/IP.    Most of the things one might want to do with data communications are necessarily going to involve the Internet at some point.   Another aspect is that there are a ton of TCP/IP cognizant hams out there but not many experts on DUAL (whatever that is).  

So I guess it is fair to say that I support TCP/IP usage on ham packet networks, at least in theory.   In practice I stubbornly keep my AX.25 packet connection to a DX Cluster up and have not experimented with ham TCP/IP.   I'm sure I could do TCP/IP, but have not come up with a good reason to expend the effort.

Here is another point, speed is everything in data networks.  My experience comes from being intensively involved in the Cellular industry's implementations of wireless data service.  We have had ~14.4 kb TCP/IP service available for several years, first through CDPD and more currently via CDMA Data.   Neither of these "products" has caught on with the general public and the recurring rationale is because they were "too slow".   So the markets for these products were limited to a few niches where speed wasn't an issue like point-of-sale terminals and text based police dispatching.   Lately we've rolled out 3G Data that offers 40 to 60 kb thruput to the end user.  And guess what, sales have mushroomed.   Why?  Because it is now fast enough to be used for everyone's favorite application, the WWW.

So my point here is that support (or lack of it) for a network is primarily a function of having an application that people find useful and delivering it at a speed that is percieved as at least "reasonably fast" by the network users.

The US AX.25 network flunked on both counts.

A better protocol is not going to have much of an impact.

Finding a meaningful application would.

DX Cluster and APRS are probably as close to useful ham apps as we are going to come up with, and they have proved insufficient in driving support for the networks.   The only thing out there with the necessary level (and breadth) of appeal would be Internet access.

Another point on which I agree with N5PVL is that wire line links have little or no place in ham networks.     This for me is mostly an emergency preparedness consideration.    A ham network that goes down in the same emergency situation that took down the commercial networks is useless for emergency purposes.    Emergency preparedness is too intimately involved in the rationale for the existance of ham radio to ignore in network design.  

In summary, the only way to save the packet networks is to make them useful to a substantial portion of the ham population.   In today's world that boils down to allowing Internet access.   All this bickering about protocols is pointless in the absense of a broadly desirable application.

The productive way to "save the packet nets" is to start lobbying the FCC to allow a useful level of Internet access from ham radio networks.   The rest will then take care of itself (and we will all have a lot of fun making it happen).

73 de Jack KZ5A  

 

</span><table border="0" align="center" width="95%" cellpadding="3" cellspacing="1"><tr><td>Quote </td></tr><tr><td id="QUOTE">Another point on which I agree with N5PVL is that wire line links have little or no place in ham networks. This for me is mostly an emergency preparedness consideration. A ham network that goes down in the same emergency situation that took down the commercial networks is useless for emergency purposes. Emergency preparedness is too intimately involved in the rationale for the existance of ham radio to ignore in network design. [/QUOTE]<span id='postcolor'>

I also agree on this, to a certain extent.

I'd love to see an all RF emergency amateur radio network, but don't think it's quite practical. Here are some examples:

1.) It's very rare for an Internet outage to effect a large area at once, thanks to dynamic routing (BGP, etc.), multiple NOC feeds, backups. A well designed network always has backups (e.g., radio backups wireline, wireline backups radio). An amateur radio network would probably be similar.

2.) Believe it or not, amateur radio spectrum is all secondary, even HF. Amateur packet radio stations near naval ports (San Diego & Norfolk for example) are routinely stomped on by military radar. My Jane's radar and EW book shows radar and comm. operating frequencies over all of HF, 2 meters, 440 MHz and 1296 MHz.

3.) Amateur packet radio stations often have more mechanical failures than commercial stations. Snow and ice around here often take out amateur radio antennas, while the phone lines and expensive commercial radio hardware continues on.

4.) The number one point in an emergency situation is to handle all traffic without errors and without arcane abbreviations (plain english) continously for many hours on end. This is where amateur packet radio really excels.

 

The most fundemental challange in establishing a nationwide RF backbone is finding a way to implement and manage a coherent, coordinated backbone topology.  

The usual VHF packet nodes are designed to allow uncoordinated implementations.   Joe ham can bring up a new NetRom etc kind of node and reasonably expect it function with only minor local coordination.   This sort of network paradigm is fine for Metropolitan Area Networks but doesn't scale up well.

There have been some interesting regional networks like Texnet in the pre Greg days, and presumably todays Flexnet implementation in the NE.   I'm not really familiar with Flexnet, but Texnet had the technical potential to expand nationally.

Whats missing in han data networking for the most part is a longhaul backbone network to tie all the various regional nets together.   To work well it would have to be coherent at level 3 and considerably faster than 300bd.  Various link level protocols and transmission methods could and should be used.   And we could design a new ham level 3 protocol, but personally I think it would be a foolish exercize in re-inventing the wheel.   TCP/IP is really the only viable candidate for level 3 in a national network.

One problem with a TCP/IP backbone is that it requires a centralized authority.   Somebody has to "own" and manage the DNS systems for instance.   Probably more or less what TAPR had in mind.   TAPR seems to lack the support required from the general ham population to make that fly.

I'm thinking that the ARRL is probably the only organization with enough breadth of support to function as a central authority.   However a nation wide ham RF data network doesn't seem to be on their agenda.  

We may have a real oppurtunity with the "Homeland Security" efforts going around to get some of the rules changes we need to make nation wide ham networking more feasable.   I'm thinking of things like the excellent suggestion further back in this thread to change HF data rules to address bandwidth rather than baud rate, might be currently doable  if associated with Homeland Security.    

Homeland Security could also be the catalyst to make a nation wide ham data net happen on an organizational level.   If HS defined the need, prehaps the ARRL or TAPR or some other group would step up the plate to provide the organizational support.   HS could even be a source of funding.  

Oh well, I'm not holding my breath while waiting for the ARRL to take a leadership position in data.   I think they are too busy trying to preserve CW and bemoaning the passing of sparkgap (which BTW is making a comeback as UWB).   Perhaps I exaggerate.

To get back to the original subject of this thread in which my bro N5PVL is extolling the virtues of 1200 baud and funky protocols, consider this.   We had a fairly well developed 1200 bd net several years ago.   It didn't die because the underlaying protocol wasn't 30% more effiecent.   It died because it wasn't useful to an adequate number of hams.

A new protocol will not make it significantly more useful.   No protocol, TCP/IP, Netflex, or otherwise will make ham packet networks popular again by itself.   You have to look at the fundemental purpose of a network which is to connect users with applications.   To have a happy network you have to have all three elements.  We have the users, we have the technical skills to produce a network, we do not have an application that appeals to a suitably broad spectrum of hams.

RF based Internet access is the only application I'm aware of that has a snowballs chance in you know where of filling the application void.   Wireless Internet access is booming in the commercial world and as hams we should be leading the charge, or at least following.   Instead we are still wondering why hams got bored with 1200 bd and arcane protocols.

Don't get me wrong, arcane protocols are wonderful if you are into developing and testing new stuff, but they are unlikely to become the basis for anything nationwide.

So yes, we need more speed and we need a RF based nationwide IP backbone network.   We are not going to get any of the above untill we get a broadly useful network application.   "Build it and they will come" has already had its day and failed, both on ham radio and in the commercial wireless data world.   The network is not and end in itself (unless you are a network experimenter or developer) it is only as useful as the applications it brings to the users.

73 de Jack KZ5A

 






 

   

 

RE:

</span><table border="0" align="center" width="95%" cellpadding="3" cellspacing="1"><tr><td>Quote </td></tr><tr><td id="QUOTE">
N3NMN's HYBRID SOLUTION:
----------------------------------
(a) Use 802.11 for USER Access!
(b) Use IP-encap on AX.25 networks for longhaul backbones! (500kb-750kb)

The technology is already out there, it just needs to be installed.

PC/FlexNet is a real nice platform to work from because NETROM's IP router frankly stinks.

So who has objections to this solution?

[/QUOTE]<span id='postcolor'>


I have an objection, about the use of 802.11 stuff for user access. Actually, I have three objections about that part of the solution you have offered, but the first objection makes the others moot, IMO.

--

First objection: It is not practical:  I just do not see 802.11 as having enough "horsepower" to handle user access.

Just as with your nodes, your users are often some distance away.

I have heard about 802.11 stuff being pushed into usefulness over decent ranges, but in every case I have heard of, this was accomplished using a pair of parabolic dish antennas, one at each end. - And as we all know, user access needs to be omnidirectional, since users tend to live in all directions.

You could split up the transmission line several times and use an array of parabolic dishes to simulate an omnidirectional pattern, I suppose, but at this point you are also looking at lots of extra expense for very little return. There will still be nulls in the pattern, and you can be sure that outraged Hams will live right in the middle of them.

Besides that, tower owners may not be delighted at the prospect of having a ring of wind-catching dishes around their tower. Most times Hams get tower space, in part, by being unobtrusive. I have never heard of a case where Hams could "do anything they like" on donated commercial tower space.

--

Second objection: It is not good networking practice: The user access is faster than the backbone.

A general rule of RF networking is that you want your backbone links to be at least eight times faster than the access to that backbone. If the backbone is faster than access, then several users can use the backbone at once transparently, that is to say without being aware of each others' presence. As more users come online at once, they become more and more aware of each other, due to the way the networks' performance progressively deteriorates.

If the access is faster than the backbone, then even one user will experience a major slowdown! Put two, or three users on there, and the term "unusable" will come up regularly in conversations about the network.

A good user access speed for a 500kb to 750kb backbone would be 62.5 to 93.75kb, using the "factor of eight" rule.

If your network is to support many applications, and/or many users though, then even the 8/1 ratio will not really be good enough.

The Texas TexNet network, for example, ran 8/1, with 9.6kb backbone and 1.2kb access. Still, TexNet (in its heyday) had serious congestion problems and for a few years there, would slow down so much as to effectively "lock up", every evening. All from Hams using and enjoying the network, something we are all hoping to see happen.

If the TexNet 9.6kb backbone links had been fullduplex, the network would have performed much better.

What ratio would you call 9.6 fulldup to 1.2 halfdup, I wonder? I know that it is several times better than eight to one.

If you slow down the 802.11 stuff enough to serve reasonably as user acces to a 500kb to 750kb backbone, you still have serious problems with range and omnidirectional coverage, but perhaps not as bad as before, at thier rated speeds. In the long run, you will probably be better off using off the shelf Ham equipment for user access. The reason for this is that at those lower speeds, lower frequencies are much easier to work with.

--

Third objection: It's not all on Ham frequencies:

From what I understand, some 802.11 stuff operates on Ham bands, and some doesn't. I have noticed that some Hams are kind of "fuzzy" on this issue, but it should be obvious to all that as Hams, we benefit when we use our special frequencies and harm our own cause when we go elsewhere *for Ham Radio communications*.

As Hams we are judged by what we do with and on the Ham bands. If we act like we are not interested in using them, they can very easily "go away".

--

So I'm all for high-speed backbone and using FlexNet so that your network can "route everything", including applications we haven't thought of yet, but I would back away from 802.11 stuff, because it is not really suitable for our purposes.

Charles,  N5PVL

 

Sorry to be gone so long... I've been working on a new packet networking web site.

USPN Website

Mainly I've been doing (and re-doing) the HTML. Check it out, leave me a note, let me know what you think.

Imagine!   A web site dedicated to amateur packet radio networkers in the U.S.!

Charles,  N5PVL