Steve S. Bosshard (NU5D) wrote:

> I don't rely on one touch reply because when folks change repeaters or
> bands it takes some time for the far end gateways to catch up. For me
> it is much better to call the repeater instead of the individual.

Since no one has probably posted this to the regular D-STAR list...

There were some updates to the U.S. Gateways recently that will help 
speed up certain types of repeater changes.

Let's set this up with a scenario:

If you're talking to me and I'm on W0CDS Port B, and you're on W0CDS Port B.

And I switch to W0CDS Port A.

Callsign routing should work virtually instantly now in that scenario. 
The local Gateways now keep a live "local routing table" for local users.

Additionally the Gateway will look in its local table for calls coming 
in from outside, I believe.

If the Trust Server hasn't "caught up" yet, the local Gateway is still 
looking at its local table when the call arrives for your particular 
callsign.

Assuming this works the way the authors intended, and your local Gateway 
got the upgrade over the weekend...

Things should route pretty fast now if someone moves repeaters, as long 
as they've a) keyed up on the other repeater, and b) their data/callsign 
wasn't mangled by a bad path, etc.

In my mind, but perhaps not in Icom's... this is just a baby step toward 
de-centralization of the lookups.  Maybe down the road, individual 
Gateways could query each other, keep their own tables, and the only 
purpose of the Trust server would be to pass IP information of where the 
Gateways can "find each other".

It seems senseless to centralize the actual lookups (the situation we 
had before this upgrade) -- that just slows things down.  On the flip 
side, Gateways querying each other raises network latency and other 
problems that could cause more individual queries to fail, so the Trust 
Server probably has to stay involved in a "future/dream" system as the 
"backup" way to look up data.  Gateways might in such a system, query 
each other first, and then look to the Trust Server as a last-ditch 
effort.

The overall problem with any lookup based system is timing... each 
trigger of the PTT is a lookup, and it has to happen relatively quickly 
for things to work in a source-routed system like the current one.

People have joked that what we're attempting to build technologically 
mimics the telecommunications SS7 or (Signaling System 7) network, 
almost to a "T".  (There's a joke hidden in there, since "T" is the 
stock ticker of a particular company that had a lot to do with creating 
SS7.)

SS7 is what looks up where to route your phone calls, every time you 
call a number your local telco switch doesn't know about.  It has to 
return an answer world-wide in just a couple hundred milliseconds. 
Adding local numbers to SS7 (instead of just toll-free numbers) is what 
led to the feature we all know today as "Local Number Portability".  The 
network no longer cares about the first three digits of the number, but 
those used to be the "destination-routing" information... those three 
numbers meant a particular switch was going to service that phone call 
-- in North America anyway.  Nowadays, any number can be serviced by any 
switch.

So... D-STAR will get there... it'll just take some time.

Meanwhile... try out the changes already made.  You can jump from one 
local repeater to another and callsign routing to you should work as 
long as you've keyed up when you switched...

If you like the new behavior, thank Brian NJ6N and Icom.  Brian wrote 
the upgrade script, Icom wrote the behavior change, as best as I can tell...

Nate WY0X

Reply via email to