[Xastir] Change to RELAY digipeat function

[Gerry Creager N5JXS]

I'm glad it's working in your area.

hasan schiers wrote:
> It's not that hard...Bob has documented it pretty well. The one thing 
> that is good is that there has been a robust discussion of what should 
> be done and how to implement it on varioius platforms.

There remains some question about his version of ALOHA vs CDMA-CD theory...

> I'm doing a massive upgrade of debian, xastir, and installing the new 
> digi-ned with the agwpe hooks today (with a LOT of help from Ryan, 
> KB0JQO). When complete, we will have implemented the new paradigm, 
> including maintaining RELAY support for a period of time, the WIDE1-1 
> alias for fill-in digi's, as well as the full WIDEn-n and Abusive Path 
> Trapping schemes, for all the platforms in our coverage area.

This *will* help: you have a more adaptable system allowing for faster 
changes when they are mandated.

> Right now we have implemented this for:
> 
> UI-View32 smart digi's
> UI-Digi (as much as one can until author updates firmware)
> KPC3+ digis
> Digi-ned (one in operation and one to go on line today)
> 
> WinAPRs is not being used in digi-service in our area, other than as 
> fill-in. There is nearly 0 WinAPRS use in Iowa. UI-View32 dominates, but 
> there are several Xastir sites as well. If things go well here today, I 
> will be converting from UI-View32 based smart digi'ing to Digi-ned based 
> today sometime.
> 
> So far, it has been very easy for two reasons:
> 
> 1. We didn't start communicating with the user base until the dust had 
> largely settled.
> 2. Two people did all the implementation/design work for the change.

These two are key for a roll-out like you're discussing here.

> We noticed between a 20 and 40 percent drop in traffic! (This is based 
> on actual traffic counts taken by the minute/10-minute/ periods and then 
> averaged over a period varying from 1 to 240 hours.) This process has 
> been ongoing for over a month and the data is clear, so instead of 
> speculating about how it "can't possibly work"...do the real work, make 
> the changes, COMMUNICATE WITH THE USER BASE, and see for yourself. Of 
> course to actually see what is taking place, one has to actually spend 
> the time to monitor the channel, then collect and analyze the data. We 
> (the hicks from the sticks) did it....anyone similarly motivated can do 
> it and not have to deal in speculation and conjecture.

Let's talk in another 2 months.  A month is a short timeline for 
monitoring traffic of this sort, IMO.

Taking the time to monitor and analyze the data *IS* the only way to 
finally get the numbers.  And then, if it is working in your area, 
that's great news!  And, no, I'm not being sarcastic.

> ...and it is NOT the result of paths being wrong...we monitor the 
> channel many hours per day and have corrected 98% of the users paths.

Again: User acceptance and change in your area may not be representative 
of the rest of the country.

> There are two real keys to the change...
> 
> 1. The network MUST be monitored on a continual basis.
> 2. The users must be instructed, educated and given feedback.

The majority of the APRS population has done a poor job of both of 
those.  That you've decided to do both is a stellar accomplishment.  I 
suspect that, if more monitoring and analysis were performed on the East 
Coast, we could see other trends evolve.  I don't think this will fix a 
lot of the congestion issues there, or in California.  I could be wrong. 
  I still expect the benefits in heavily congested areas to be transient.

> In other words, the local network must be MANAGED...gee, what a concept!

If you'll look back, you'll find I have advocated that in the past.

> We have superb cooperation among digi owners and end users. Perhaps it's 
> because we don't jerk them around and we make any changes easy to 
> understand by implementing changes "by the numbers", including 
> instructions and help. Then again, perhaps it's because we are in Iowa, 
> where people actually value cooperation over competition, and the good 
> of the network over any one individual's ego.

Could you export some of that toward the coasts?  We have a fair bit of 
apathy now from a lot of digi owners here.  We'll see the changes 
eventually implemented, or not, on that basis.  We've one group that's a 
very proactive group here in Texas.  I've still not gotten a feel for 
how they're going to go with this proposal.

> Whatever the reason(s), the change is solidly in place on all the 
> platforms and we have a significant reduction in channel clutter as a 
> result.

So: We've started patching a system that has reached the breaking point: 
Is it now time to start designing a real follow on system rather than 
reacting to problems?  That's been the issue in the past.  Bob doesn't 
want to "allow" anyone to go beyond his view of APRS, and will ridicule 
and deride anyone who wants to take a different view than his, including 
folks better than me in network theory.

> Until new hardware and a brand new scheme to make use of the hardware 
> come available, this is the best possible way to go. We did it and are 
> pleased.

And that's what matters at this point.  But the new scheme and new 
hardware need to be designed proactively instead of reactively and 
adapting existing hardware.

73, gerry

> ----- Original Message ----- From: "Steve Friis" <wm5z at comcast.net>
> To: "James Ewen" <jewen at shaw.ca>
> Cc: <nwaprssig at nwaprs.org>; <xastir at xastir.org>
> Sent: Saturday, April 09, 2005 2:11 AM
> Subject: Re: [Xastir] Change to RELAY digipeat function
> 
>> Thanks to all that have been part of this discussion. I am listening 
>> to all that is said because here in Las Cruces, NM there doesn't seem 
>> to be very good coverage. I am thinking of seting up a digipeater 
>> here, and am learning of the do's and don't's. See, some of us are 
>> willing to do it right, if only we knew what was right.
>>
>> Steve/WM5Z
>>
>> James Ewen wrote:
>>
>>>> #2, but proceed with caution after comprehensive testing and
>>>> re-testing, to make sure we're not fooling ourselves.
>>>
>>> That's the hard part. To be able to do comprehensive testing, you 
>>> really need to make wholesale changes to all of the parts... You can 
>>> change 1 digipeater, but it won't make a difference if the rest of 
>>> the digipeaters still implement the old parameters. The digipeaters 
>>> need to drop RELAY, and home stations need to change their RELAY into 
>>> WIDE1-1 (only if they are in a useful area). The trackers (meaning 
>>> all APRS stations, mobile (whether full blown or dumb), or home 
>>> stations) need to use WIDE2-2 or less, and if absolutely necessary, 
>>> WIDE1-1,WIDE2-1 to get a boost from the home stations.
>>>
>>> Everything needs to make the change to the new parameters to be able 
>>> to make proper quantitative and qualitative analysis of the 
>>> implications of said changes.

-- 
Gerry Creager -- gerry.creager at tamu.edu
Texas Mesonet -- AATLT, Texas A&M University	
Cell: 979.229.5301 Office: 979.458.4020 FAX: 979.847.8578
Page: 979.228.0173
Office: 903A Eller Bldg, TAMU, College Station, TX 77843