craig@gpu.utcs.utoronto.ca (Craig Hubley) (06/15/89)
I realize there are several people collecting or posting reviews of good introductory books on the OSI protocols, but do any of these references touch on interconnections with ISDN, or at least make it crystal-clear where ISDN's standards are supposed to interconnect with OSI's? It seems to me to be a straightforward question, but I have yet to hear a straight- forward answer. I am interested specifically in connecting private OSI LANs to common-carrier ISDN WANs. I have *no interest* in implementation details save those directly relevant to application design. If there are no such books, I strongly suggest someone write one. :-) In the abscence of such references, is there a general overview that makes the responsibilities of the various standards groups clear? I have yet to see a document where this is outlined. X/Open, OSF, and the like seem to have confused the issue by introducing private consortia into the picture. In the meantime, is there a mailing-list or newsgroup source of ISDN application design information ? I apologize in advance if this question has been brought up before. Please send replies (and flames) by email only, Craig Hubley -- Craig Hubley ------------------------------------- craig@gpu.utcs.toronto.edu "Lead, follow, or get out of the way" mnetor!utgpu!craig@uunet.UU.NET ------------------------------------- {allegra,bnr-vpa,cbosgd,decvax,ihnp4,mnetor,utzoo,utcsri}!utgpu!craig
levin@milk10.uucp (Larry Levin) (06/16/89)
In article <1989Jun15.160207.18925@gpu.utcs.utoronto.ca> craig@gpu.utcs.utoronto.ca (Craig Hubley) writes: >...... I am interested specifically in connecting private OSI LANs >to common-carrier ISDN WANs. I have *no interest* in implementation details >save those directly relevant to application design. > Frankly I do not think that ISDN is an acceptable bridge media between LANs. The problem is lack of bandwidth in ISDN (as opposed to B-ISDN, the broadband standard) as well as regulatory issues, although the regulations may differ outside the US. The first question you need to ask is if a single 64 Kbps B channel is sufficient for your inter-LAN traffic. If all you send is E-mail the answer is probably yes. If you try to run a diskless workstation off a file server the answer is no. In theory you could use multiple B channels for a single bridge. This can be done but the bridge has to transfer each individual packet over a specific link since splitting the packet may result in corruption due to differing transmission delays on each channel. I dont know of any bridges on the market to this at this time. I do recall a company in the Chicago area that sells an "ISDN" bridge. the target application in this case is to allow a single PC or workstation to use a single B channel as if it was an extended AUI drop-cable. As to the regulatory side, take a look at the tarif for T-1 versus a Basic or Primary ISDN circuit. I admit I dont know the numbers but I here that T-1 is considered a bargin these days. Two final points: In regards to you original question on OSI vs ISDN, see "ISDN Technology, Concepts, and Applications" by P.J. Kuehn, Springer-Verlang Lecture notes in Computer Science, Vol 248. In regards my response, I expect a number of folks on the net, especially at the BOCs and Bell Labs have actually tested some of these assumptions. If they have opinions or experience contrary to mine I look forward to reading about it here. Larry Levin SRI International ---- My opinions, love em or leave em ---
goldstein@delni.dec.com (06/16/89)
In article <1989Jun15.160207.18925@gpu.utcs.utoronto.ca> craig@gpu.utcs.utoronto.ca (Craig Hubley) writes: >...... I am interested specifically in connecting private OSI LANs >to common-carrier ISDN WANs. I have *no interest* in implementation details >save those directly relevant to application design. The Corporation for Open Systems (COS) has a COS Profile Selection which includes ISDN in some of its profiles. While these are still in draft form (I personally question the need for ISO 8208 below 8473 and above 7776 in circuit mode...), the profiles do illustrate ways to use ISDN in an OSI context. ISDN provides packet mode service up to Layer 3 (CONS or CO-subnet) as well as circuit-mode service (switched L1) at various speeds, so long as they're 64 kbps. :-) The standards also define circuit-switched H channels at 384, 1472, 1536, and 1920 kbps, but I don't think those are close to implementation in any of your common central office switches. Their applicability to any given application of course depends upon bandwidth requirements. Applications, of course, should be sufficiently isolated from lower layers that it shouldn't matter whether you're on an ISDN or what-have-you, other than bandwith, speed and the other usual WAN issues.
schaen%bland@GATEWAY.MITRE.ORG (Sam Schaen x7297) (07/22/89)
Concerning your request for a reference that discusses the interconnection of OSI LANs to ISDN WANs: There is a brief section on ISDN protocols in Stallings "Data and Computer Communications", second edition. Not a great deal of detail, but it's a start. I think one of the problems you're seeing is that the ISDN implementers have concentrated more on the telephony side until recently. I understand that's beginning to change, but I'm not following the situation very closely. Sam Schaen MITRE Corp.
dcrocker@AHWAHNEE.STANFORD.EDU (Dave Crocker) (07/22/89)
Let me try to parade my ignorance and see if anyone wants to join, waiving flags are playing appropriate marching music: As I perceive ISDN, it is a conduit for raw bits, roughly on the "level" with X.25. (Lots of differences in mechanism, etc, but its actual utility to a person/resource is at that level.) Hence, it fits into OSI Layer 2 and the sub-network sub-layer of Layer 3. To plug it into the rest of the OSI stack, you build an appropriate Subnetwork Dependent Convergence Module, also in Layer 3, and then the rest of the stack doesn't know there is anything special to deal with. Yes? Dave
frg@jfcl.dec.com (Fred R. Goldstein) (08/03/89)
In article <8906172235.AA25320@ucbvax.Berkeley.EDU> dcrocker@AHWAHNEE.STANFORD.EDU (Dave Crocker) writes: >Let me try to parade my ignorance and see if anyone wants to join, >waiving flags are playing appropriate marching music: > >As I perceive ISDN, it is a conduit for raw bits, roughly on the "level" >with X.25. (Lots of differences in mechanism, etc, but its actual >utility to a person/resource is at that level.) Actually, ISDN Packet Mode (X.31) is pretty much identical with X.25 if you use one of the options (circuit-mode call to an X.25 switch); if you use the other option (packet-mde call), the data transfer phase is still pretty much the same thing. ISDN didn't reinvent the wheel. >Hence, it fits into OSI Layer 2 and the sub-network sub-layer of Layer >3. To plug it into the rest of the OSI stack, you build an appropriate >Subnetwork Dependent Convergence Module, also in Layer 3, and then the >rest of the stack doesn't know there is anything special to deal with. > >Yes? Yes! Indeed your description is remarkably accurate, especially since most ISDN weenies don't know what the subnetwork role (not "sub-layer", for some odd reason ISO doesn't like that term) even is. But ISDN packet mode is precisely a connection-oriented subnetwork that is capable of delivering the OSI Network Service, per ISO 8648. ISDN Circuit Mode is just a Layer 1 bit-delivery service. fred (member, ANSI T1S1)