Hi Valdis, Thanks for the reply. SMPP is a good example to what I am looking for, and I am trying to see if other protocol foreseen in the future (or past) have the same characteristics: The special thing about SMPP is that in the same TCP/IP session packet that are really SMS embedded in IP are sent from one network element (called SMSC) to another (called ESME) - so that the packets (i.e. SMS) belong to different mobile users. In this case I cannot map the IP address of the TCP session to one specific mobile subscriber - and the only way I can identify the subscriber is by "looking" on the SMPP layer (above the TCP) and extract the subscriber mobile number. Another example may be SMTP (not sure). Anyway the thing I am trying to solve is whether I should consider this kind of protocol characteristic as "special" and "rare", and so build a dedicated solution for SMPP, or maybe I can expect to see many more such protocol and so I should build a more generic solution. It is obvious to me that a generic solution is better technically, but since my resources (i.e. time) is limited, I need to see how rare this kind of protocol (same TCP/IP with packets of multiple users) is common/expected. Another characteristic of SMPP that I am looking for its "commonness" is the fact that there could be many TCP/IP session from many SCSCs to ESMEs at the same time - and packets of the SAME subscriber may appear on them at the same time... Complicated :-) that's my life (-: yours, Haim Rochberger > -----Original Message----- > From: Valdis.Kletnieks@xxxxxx [mailto:Valdis.Kletnieks@xxxxxx] > Sent: Sun, January 25, 2004 6:43 PM > To: Haim Rochberger > Cc: 'ietf@xxxxxxxx' > Subject: Re: packets of multiple users sent over the same > TCP/IP session > > > > On Sun, 25 Jan 2004 15:42:28 +0200, Haim Rochberger > <haimro@xxxxxxxxxx> said: > > > I will appreciate your help as it will help in making a > better design > > decisions. > > Hmm.. been a while since I've had to post this. ;) > > The first thing you want to do is take a step *back* from the > problem, and ask > yourself what problem you're trying to solve via this > multiplexing. You > already said you didn't want to use tunneling, but at the > same time you gave us > insufficient information to know if SMPP or anything similar > is applicable to > your problem. With the information you gave, it's likely many > will fear to give > you any really useful answers for fear of sending you in a > totally useless > direction. > > For all I know, HTTP pipelining may be a solution to your > problem - in that > scenario, you have in-flight packets for multiple HTTP requests (read > "subscribers") active at the same time. On the other hand, > although it meets > the requirements you stated, it's almost certainly not the > right answer. But > not knowing what problem you're trying to solve, it's hard to say. > > Basically, you need to make sure that what you have in your > hand is a nail > rather than a screw *before* you start selecting hammers.... > >