Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!mnetor!seismo!mcvax!diku!olamb!soren From: soren@olamb.UUCP (Soeren Rabbe) Newsgroups: net.dcom Subject: Re:Re: Networking on UNIX - 3 approaches (really OSI Layers) Message-ID: <202@olamb.UUCP> Date: Mon, 25-Aug-86 16:32:33 EDT Article-I.D.: olamb.202 Posted: Mon Aug 25 16:32:33 1986 Date-Received: Mon, 25-Aug-86 22:05:59 EDT References: <964@hoptoad.uucp> <3535@amdahl.UUCP> <1019@hoptoad.uucp> Organization: AmbraSoft, Copenhagen Denmark Lines: 67 Keywords: networks protosw LAN OSI Without interfering with the discussion on UNIX(R) streams, I would like to give some comments on the part of the discussion covering the OSI Layers. In article <1019@hoptoad.uucp> tim@hoptoad.uucp (Tim Maroney) writes: >In article <3535@amdahl.UUCP> sjl@amdahl.UUCP (Steve Langdon) writes: >>In article <964@hoptoad.uucp> tim@hoptoad.uucp (Tim Maroney) writes: >> . . . >>I should add that my praise for this >>example does not mean that it is necessary, or desirable, to view all OSI >>layer interfaces as software module interfaces. Efficiency will often >>require layers to be implemented without an exposed "service" interface. > >I could not possibly disagree with you more. If you have not provided a >software interface for a protocol except one usable by higher-level >protocols in the same project, then you have not provided the protocol. If >a protocol is designed in such a way that no efficient software interface >can be provided, then the protocol has been very poorly designed. It is my impression that some clearifications can be needed on this point. All "lower-level" protocols are made to be usable *only* for the higher-level protocols, i.e., layered protocols are made to structure the functions of the transmission service, not to exist by them self. What all the protocols in a layered model ends up with is a service provided by the highest lay- er - in OSI the Application Layer. But this does not imply that the service interface must follow a standard or even exist within a system to say that the system follows a given architecture or, if the protocols implemented are "OSI protocols", to say that the system is a Open System. In the OSI Basic Reference Model (ISO International Standard 7498) it is said (Sec. 6.2 Note 2): "2 It is important to note that OSI *per se* does not require interfaces within open systems to be standardized. Moreover, whenever standards for such interfaces are defined, adherence to such international interface standards can in no way be considered as a condition of openness." This is the same as saying that the "openness" of a system is only determined by the protocol interface and not by the internal service interfaces. Then, if the system by itself does not include all layers or it is the intention that users should have the possibility to add there own functions; e.g., special higher layers, it would be very bad design not to provide at least some very well define service inter- faces internal in the system. I hope these points can be helpfull in future discussions on this topic. I have a weak point when the discussions takes the OSI Basic Reference Model and the model introduced by the standard as a "Bibel" without identifying the scope of it [:-)]. N.B.: The opinion expressed hereby is purely on the behalf of the undersigned and have no connection to my membership of the Danish OSI Standard Committee (equals ISO TC97/SC21). -- ---------------------------------------------- Soren Rabbe AmbraSoft A/S, Denmark soren@olamb.UUCP