I've received the following via Zheng Wang at UCL, regarding PIP. Anyone interested, please contact Zheng directly at Z.Wang@cs.ucl.ac.uk. Tim ----------------------------------- * From: francis@com.bellcore.thumper (Paul Francis \(formerly Paul Tsuchiya) * To: j.crowcroft@uk.ac.ucl.cs * Subject: installing pip....? * Cc: z.wang@uk.ac.ucl.cs * * Hi Jon, * * We want to install a number of Pip hosts and routers * around the internet so that we can 1) show a nice * map of Pip infrastructure in Amsterdam, and 2) genuinely * start playing with transition. * * I'm assuming that you can bring up one or two Pip hosts * like you did before, but I'd like to ask if you have a * few friends in UK or The Continent that would also care * to participate..... * * The following outlines what we have in mind...... * * Thanks, * * PX * * ________ * * * We are interested in starting to populate the IP * internet with a number of Pip systems. This will * allow us to start experimenting with various aspects * of Pip, particularly with transition from IP to Pip. * * Pip is one of the contenders for replacement of IP. * IP is suffering from lack of addresses and scaling * problems in routing. Pip's virtues are that it has * new routing features such as provider selection, is * easier to administer than IP or the other proposals, * switches fast, and can more easily evolve over time * to new functionality. This evolvability will allow * us to install real-time flows as they are developed. * * To read about Pip, see the May 1993 issue of IEEE * Network magazine. You can retrieve that article * from thumper.bellcore.com:pub/tsuchiya/pip/ieeeNet.ps. * Additional reading can be found in the internet draft * repositories, especially draft-ietf-pip-processing-01.txt * and draft-ietf-pip-architecture-00.txt. * * We are initially looking for a small number of "friendly" * sites for initial experimentation. As we gain experience, * we will be open to anyone who wants to participate. * * Pip transition involves creating a logical Pip * infrastructure by tunneling over IP. It also involves * translating between Pip and IP, for when a Pip hosts * talk with IP hosts. Thus, your Pip host will be able * to communicate with any IP host--you won't even know that * it is a Pip host (unless you want to play with its * provider selection feature). The tunneling puts Pip * over IP with IP protocol number 4 (which means IP over * IP). Thus, your border gateway filtering will have to allow * these packets to pass through. * * The main purpose of the initial experiment is to play * with our configuration software and make sure that we * can successfully create a Pip infrastructure. * * Our software runs on Sun Sparcs running SunOS 4.1.2. * We are flexible as to the extent of your involvement. If * you don't want to do anything, but can supply us with the * super-user password for your machine, we can do all of * the installation from here. Otherwise, we can supply you * with the binary, configuration files, and installation * instructions. If you elect to install it, then you can * be expected to have to reboot the machine and bring in * new configuration files from time to time. We can * probably expect a couple of false starts, so you may have * to reboot several times in the first couple of weeks. * If possible, the super-user method is preferable to us. * * We hope to start installation around June 15. Before we * can start, we need to get commitments from participants, * design the logical topology, and generate the configuration * files. We can expect to start expanding the experiment * around August, and adding dynamic routing (thus eliminating * the need for static configuration) in October.