fey@dccs.upenn.edu (Mark L. Fey) (02/25/88)
Like many of you now operating campus wide networks, the Data Communications and Computing Services organization of the University of Pennsylvania is wrestling with the administration, reliability, and support issues surrounding DECnet (mostly) protocol packets crossing the campus as we sub-divide the network into IP sub-domains. Very early on we recognized the need to identify protocol standards and like many of you we issued formal statements that the DOD protocol suite would be the *ONE*. The installation of the ISN provided a convenient means to intelligently N-way bridge an ethernet through every floor of the 116 buildings of our West Philadelphia campus. This interconnection method of course did not provide the Protocol Police (P.P.) with much to do. Everything went to where ever it was directed to go and nobody alerted the P.P. The P.P. grew fat and lazy; then the need to install IP routers to further enhance the reliability of the network and subnet the campus to minimize traffic in an administered manner became evident. All those folks doing their own DECnet thing suddenly need *REAL* support as opposed to benign neglect. This problem of dual networks (virtual or physical) or restriction to subnet DECnet access is a very general problem with all similarly constructed networks. Some of you have faced this immediacy in the recent past, some are currently facing it and some are ducking this issue as long as you can. What I need to do is to talk to and perhaps visit sites that allow DECnet packets on the same ethernet as the TCP/IP ethernet and use P4200 or Cisco routers with DECnet support turned on as well as conduct the same discussion and site visits with those of you who support two separate inter-subnet networks. Both of these groups have different technical and administrative problems. Staffing of the network support could be different in each case with regard to the skill set each engineer and technician must possess. The cost alternatives must somehow reflect the choices that are made. Of course, I will be happy to summarize to the net the feedback this article generates as well as the subsequent results of dual protocol decision making at PENN, since I recognize the wide interest in this topic. Both Campus networks and national networks have struggled and are currently struggling with these issues. ******************************************************************** Mark L. Fey fey@dccs.UPENN.EDU Program Manager, (215) 898-1145 Data Communications and Computing Services Suite 221 A University of Pennsylvania 3401 Walnut Street Philadelphia, PA 19104 ********************************************************************