Bonnie, I have seen that Rein, the designer of the system, will reply to your questions. I will let him do so since he has more knowledge than I have on the PSKmail system.
I might comment after on the potential integration of the concepts for HFN since I have also used this system under Multipsk. 73s, John (VK2ETA) --- In digitalradio@yahoogroups.com, "expeditionradio" <[EMAIL PROTECTED]> wrote: > > Hi John, > > Thanks very much, for the detailed comments on > PSKmail for this type of application. That is the > probably the closest I've seen to approaching > "push message" capability. > > Here's some follow-up questions: > How does the mobile operator determine which > PSKmail base and frequency to check in to > at any particular moment? Is it manually > selected by the operator, i.e., does the mobile > operator need to keep manually checking if they > remain "linked-to-base" on an ongoing basis? > Does the op need to keep finding another base > to be linked-to-base" with, so that the notification > messages get routed properly? What is the fall-back > position for the system for notification, in the > case that none of the bases show a link-to-base > condition when the target mobile op hasn't checked in > for a while? Is there a time-out or "not-linked" > indication to the network or mobile op? > > I'm very interested in the network server side of > how this can work smoothly. I think that it is the > key to getting the best notification system. It > would make sense to join as many ham networks as > possible to this, to enable a message to be routed > to the target operator by any method they are using. > A mutual cooperation between ham networks could > be forged, and this could make it a reality. > If one dials a telephone number, it isn't necessary > to know which telephone provider company that the > called party is using. We need to carry this type > of universal networking into the ham radio realm. > The "email address" provides universal portability > and networking opportunities for hams. Hams could > adopt specific email addresses that are used for > emcomm purposes, and use email forwarding. This > simple feature could be leveraged to provide > powerful networking for hams. > > At least for the HFN system, the high probability > of linking on HF has already been achieved through > the power of a network of geographically distributed > HF base stations running simultaneously on all bands. > > It has often been pointed out that HF base-to-mobile > can be statistically undependable for 24/7 point-to-point > communications with a base station, due to the changing > ionospheric propagation and channel conditions. > > However, the statistical probability for successful > communications with a mobile goes way up when > dynamic linking with any base in a geographically > distributed HF network is added. > > A single ham band may not be open at any given > instant between 2 specific stations. > In fact, there might not be any HF band open between > those 2 specific stations :) > > But, it is very rare that all HF bands are closed to > everywhere. > > That points to the need to develop a wider, more > flexible, network outlook for HF Emcomm systems; > one that is not concentrated so much on NVIS or > specific regular propagation patterns. > > The solar flare that happened during the Katrina > disaster response certainly taught us how fragile > traditional Net-Control-centric state nets can be > that rely on 75 meters only. > > Bonnie VR2/KQ6XA > > > John (VK2ETA) wrote: > > Bonnie, > > The way Pskmail addresses the push messages is by using two concepts: > > 1. A notion of Linked-to-a-base status and > > 2. a centrally accessible (over the internet) > > database so that servers can be coordinated and > > avoid duplication). > > > > More details: The clients (mobile units) have to > > "check in" by sending a "link to base". From there on, > > the server will push any new APRS messages > > (without ack) to the client. Of course other message > > sources could be pushed as well. >