[Top] [All Lists]

Re: Simple Network Paging Protocol

1993-09-01 13:54:12
1> It seems overly complicated. If you only allow one MESS and one
   PAGE for each send, you shold collapse it into one verb, e.g.

      ->      PAGE 5551212 Your network is hosed
      <-      250 Page Sent
      ->      QUIT
      <-      221 Goodbye!

   I helped build a pager system with a very similar protocol. Don't
   needlessly multiply entities.

2> You should allow for multiple pager IDs per message.

Funny you should mention that :-)  The reason that I did it with the
PAGEr/MESSage format was so that eventually, hopefully, I could
figure out a "good" way to provide for multiple pager IDs per
message.  As you are aware, IXO doesn't really tell you that a
pager ID *was* invalid until you have assembled the whole message and
sent it to the paging terminal.  Big problem.  This means that you
can't really validate a pager ID without trying to send it a page.

There are standard protocols in the paging industry that will allow
you to do a validation query.  That is why I've started building this
protocol this way.  Hopefully, this philosophy will allow for instant
validation of each page in a multi-recipient message.  But right now
its just a "stepping stone" to get a terminal up in an easy-to-access

3> Meaningful error responses should be numeric, with arbitrary
   ASCII strings appended for human-readability.

I patterned this after the (yecch) POP2 protocol because of being able
to report success or failure (in various degrees) by looking at the
first character of the reply.  I have, however, considered numeric
error responses in later implementations.

4> Some form of authentication should be provided.

Yeah, yeah, I know :-)   I will probably do some sort of lookup and
logging of the sender at the gateway, but is this really part of the
protocol?  I mean, look at what they deal with now: dialup phone lines
into a modem.  Thats pretty anonymous.  Anything we do on the net is
going to be more "secure" than that.

It is interesting to note that IXO does provide for a "password" to
access the IXO terminal, but nobody really uses it (rather I should
say "not that I know of") at public paging companies.

That aside, I like your approach,and this is something that sure could
use standardization!

Thanks!  This has been one of those "gonna-do's" that I have wanted
to do for some time.  My goal was (1) simplicity, and (2) something that
would be buildable for the future.  

Is this really appropriate for ietf-822?

Dave Crocker threatened to have me thrown into an industrial eggbeater
and turned into an omlette if I didn't :-)     If this is not appropriate
for this group, I will certainly refrain.  Please let me know.

Allen Gwinn