Path: utzoo!utgpu!water!watmath!clyde!att!mtuxo!mtgzy!mtgzz!drutx!druco!stanwyck From: stanwyck@druco.ATT.COM (D. Stanwyck) Newsgroups: comp.protocols.iso Subject: Re: ISDN, Layer 3 Message-ID: <3114@druco.ATT.COM> Date: 31 Aug 88 20:16:42 GMT References: <4183@pdn.UUCP> Organization: AT&T, Denver, CO Lines: 89 in article <4183@pdn.UUCP>, larry@pdn.UUCP (Larry Swift) says: > In article <3080@druco.ATT.COM> stanwyck@druco.ATT.COM (D. Stanwyck) writes: >>User-user signalling is a service where the two "users" (hint- user <> >>end-system) may send information between themselves in the call control >>messages. This comes is several varieties - some associated with a >>circuit-switched call, and at least one not associated with any other >>communication path. This latter is the type you mention above. >> >>For this connection type, the network is told that all you want is to relay >>messages between the users without establishing a related communication >>path for e.g., a voice call. You could (correctly) call this using the >>CCS7 network or whatever the network's internal common channel signalling >>fabric is based on as a packet-switching network. > > This discription almost sounds synonymous with OSI's "connectionless" > service, which isn't flow-controlled. Is that a reasonable association? Not quite. There are several characteristics which are similar to the connectionless service - yet the user-user service is connection oriented. By definition, all connection data units are self-routing - that is, they contain the complete address of the destination. User-User messages flow over a connection which has been requested by a SETUP message, and is broken by a RELEASE message. While the individual data units are neither ordered, error protected, or acknowledged, they are still associated with a connection. >>.... The brief answer is that Disconnect Request does not require the >>termination of all communication - i.e., you could potentially keep some >>type of user-user signalling up, while release request occurs some messages >>later and implies a total release of communications. > > Can you give some idea of application? Let me explain a little more. Suppose I have a voice call controlled by the out-of-band signalling mechanism. Suppose I want to release the connection with the other user (i.e., drop the voice circuit) but retain the signalling connection to the network for further use. An example of the further use that was used at the time the message set was selected (though the service is not yet defined) was the ability to have a service that would have the network deliver billing information (e.g., time and charges) after the call was dropped. The signalling order would be as follows: user sends DISCONNECT to the network, with a Facility information element requesting the time and charges facility. The network drops the voice circuit (clears toward the other end) but retains the call reference. Later, using that call reference, the network delivers the time and charges information to the user, either in the RELEASE message (terminating further signalling) or in a separate message. If the latter, then the user responds with the RELEASE message, etc. Does this help? >>Overlap sending/receiving and overlap modes relate to whether the address >>information is all carried in one message or is sent piecemeal to/from the >>exchange. Both are possibilities. > > Just curious here. Do you have an idea as to how the term "overlap" came > to be used? Overlap was used to indicate that (in the old days) the control and signalling overlapped each other. For instance, the user dial pulses a number toward the network. Let's say that number is 303-987-6543 (fictional number). As the local exchange receives the pulses, it would process the 303, determine the trunk to connect to, and start outpulsing on that trunk overlapping with the incoming pulses. Today most switches collect the entire number most of the time prior to trunk selection, but some networks (e.g., the Fed. Rep. of Germany) still have many that operate the other way. >>You really should get a copy of the current draft proposed American National >>Standard (DpANS) based on Q.931. It should be out for ballot by the end of >>this year. > > I have Q.931; is the ANSI draft substantially different? more readable? > I'll try to get a copy. Substantially different? Well, not much has been added, but some things have been deleted (e.g., suspend/resume). A few things have been changed, but overall, it is similar. Readablility - much better, as it is set as a single document (actually, two documents) by modern american word processing systems with laser printers. Not that the "american" makes it better, but the uniformity and print quality does. > Larry Swift UUCP: {peora,uunet}!pdn!larry -- AT&T o o 303-538-5004 Don Stanwyck || druco!stanwyck Denver, CO USA \__/ Telecom Standards