Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!lll-winken!brutus.cs.uiuc.edu!psuvax1!psuvm!PUCC!GETTES From: GETTES@PUCC.BITNET (Michael R. Gettes) Newsgroups: bit.listserv.nodmgt-l Subject: (no subject given) Message-ID: Date: 13 Feb 90 02:06:00 GMT Sender: Node Management Discussion Reply-To: Node Management Discussion Organization: Princeton University, CIT Network Systems Lines: 31 Approved: NETNEWS@PSUVM Gateway In-Reply-To: Message of Mon, 12 Feb 90 19:07:29 EST from On Mon, 12 Feb 90 19:07:29 EST Roger Watt said: >You missed my point, /mgr. I see nothing wrong with the idea. I merely do >not understand the rationale for creating a burdened set of 14 core >nodes. EVERY NODE THAT CAN should be running NJE-over-IP ... hundreds of >nodes, not tens! > >My question was aimed at the folk who prepared the proposal. Can anyone >at PUCC who was involved in this explain "Why not hundreds of nodes >running NJE-over-IP, each talking directly to each other?" What technical >fears did/do you have that caused you to limit the concept in this >fashion? Again, the proposal is not designed as the solution to all problems present and future. VMNET/BITNETII is not a solution to all problems present and future. The proposal was designed to deal with current problems on the network with minimal impact. Using hundreds of nodes would have had serious impact on the network. Additionally, you do not get these BITNETII connections for free. TCP circuits do occupy certain computing resources that make hundreds of connections a burden for any system. We have developed this proposal based on practical experience from the BITNETII project. The number of regions and connections was based on our projections of what could be easily handled by the network over the next year (more or less). You also have to consider some basic problems such as routing (note the discussion regarding pathalias and GENROUTS of late on this list). Lastly, I believe it would be a mistake to put your entire body into the bathtub without letting your foot in first for a while to see what it is like. Maybe we will have a complete mesh of hundreds of nodes in the future -- anything is possible. /mrg