PRIDE Progress Report February - May, 1994 Tony Bates Daniel Karrenberg 1. Management Summary During the reporting period two major milestones were reached: The PRIDE Guide 1 was released and the PRIDE Course 1 was completed. The first course has been held and was received very well. Also two revisions of the PRIDE Tools 1 have been released. These include the new tool prpath. The guide and course milestones were reached later than planned. This is caused partly by the slower start of the project as reported earlier and partly by the need to accom- modate classless inter-domain routing (CIDR) in the project. Since CIDR influences routing policy, the routing registry needs to be adapted for it. Considerable unforeseen effort of the project team has gone into speeding up revision of the routing registry specification. We decided to do this because it would be useless to continue developing tools for a classful registry. US NSF sponsored efforts in the RR area were also starting during this period. Coordination with these also needed some unforeseen resources. Fortunately the NSF sponsored "Routing Arbiter" decided to base its routing registry on the RIPE/PRIDE software. This is an important step towards a globally compatible system of RRs. Due to these developments the project plan has been revised. The project will be extended with two months recovering the 4 FTE months lost in 1993. We now plan to provide the origi- nally planned tools with exception of "prconn" in their classless versions as well as incorporate the classless routing registry into the PRIDE guide 2 and PRIDE course 2. 2. The CIDR Challenge The deployment of classless inter-domain routing (CIDR) in the Internet has now started in earnest. More and more classless routes are being originated into the Internet May, 1994 Page 1 PRIDE Report 4 pride-rep4.txt routing mesh. Aggregate routes can potentially summarise routing information which was previously used to make policy decisions with a finer granularity. This has immediate con- sequences for the routing registry as well as the tools based on it. The ripe-81 based routing registry needs to be enhanced. Unfortunately, the need to accommodate classless routing as well as its impact on the routing registry and tools were not foreseen during the definition of the PRIDE project. Consequently the work plan needs to be adapted as described below. As an immediate measure we have postponed further development of tools and helped substantially with the revi- sion of the routing registry itself. 3. Routing Registry Changes During the reporting period the RIPE routing group has taken up the challenge to adapt the RIPE routing registry to classless routing. The PRIDE project team has provided sub- stantial support for this effort which resulted in a draft routing registry specification ("ripe-81++") presented at the 18th RIPE meeting. At the same time Merit Inc. have started their routing reg- istry activities based on ripe-81. As a basis for their ini- tial routing registry implementation Merit used the RIPE routing registry software. The PRIDE team gave significant support in helping Merit to establish its routing registry. It is expected that this registry will be used by the "Rout- ing Arbiter" sponsored by the US National Science Founda- tion. We understand that the registry function of the "Routing Arbiter" will be provided to all North American service providers independently of whether they attach to the NSF sponsored NAPs or not. In order to achieve a globally compatible routing registry, Merit have been invited to participate in the ripe-81++ dis- cussions of the RIPE routing group with the aim to arrive at one RR specification both for the RIPE RR and the Merit RR. This is a significant step towards global compatibility of routing registries in the Internet. During the 18th RIPE meeting all the extensions were dis- cussed at length and a draft RR specification is expected to be agreed in early July. The current draft can be found as: ftp.ripe.net:ripe/drafts/ripe-81++.{ps,txt} A presentation of the extensions given at the ripe meeting can be found as:: ftp.ripe.net:ripe/presentations/ripe-m18-tony-RIPE81++.ps.Z Page 2 May, 1994 PRIDE Report 4 pride-rep4.txt 4. PRIDE Activities 4.1. PRIDE Tools 1 Release Two releases were made during this period, pride-tools-1.0.3 and pride-tools-1.0.4. The most significant additions to the tools was the first release of prpath. Prpath was in alpha test for quite some time. The reaction to prpath has been very good with only minor changes needed The "-g" flag was also added to prtraceroute. This allows prtraceroute to take advantage of the traceroute gateway function with IP Loose Source Record Routes especially useful for checking symmetry of policy. The current tools release is available from: ftp.ripe.net:pride/tools/pride-tools-1.0.4.tar.Z Besides the complete package, each tool is also available separately in this directory. Preliminary work has begun on a number of new modules for the tools in the form of a set of general library routines. This is still in its early stages but this is expected to be significant part of the second release of the tools package. The reaction for the tools continues to be very positive. Table 1 details the current usage statistics of the tools and the interactive PRIDE menu. The following statistics are taken from the period of February 1st - May 31st, 1994. It should be noted that this is a larger reporting period than in previous reports. Tool Ftp'd S/W(1) whois queries(2) Menu Usage --------------------------------------------------------------- Prcheck 27 695 92 Prtraceroute 79 16399 348 Prpath(5) 3 342 5 Full Package(3) 197 249(4) (1) Amount of anonymous ftp transfers of the software (2) No. of RR queries related to the tools. The average amount for each tools is for prcheck, 5 and for prtraceroute and prpath, 10. (3) Means the full release package. (4) Total number of calls to the menu. (5) Prpath was only released in the month of May. May, 1994 Page 3 PRIDE Report 4 pride-rep4.txt Table 1: Statistics for PRIDE-TOOLS-1 From this we see an increase in the amount of sites pulling the software with a decrease in the amount of queries to the RR itself from the tools. This is an indication that more sites are running there own versions of the RR software locally. 4.2. PRIDE GUIDE 1 (PRIDE-3) The first PRIDE guide was released as a draft in week 13. It followed the structure outlined in report 3. The first offi- cial guide was ready in time to be used as part of the first PRIDE course (see below). The guide is now available from ftp.ripe.net:pride/docs/guide-1.0.ps.tar.Z ftp.ripe.net:pride/docs/guide-1.0.txt.tar.Z This distribution includes a small script to make and print the document correctly. The guide is currently some 130 pages. The postscript version of the guide is the preferred format because it includes many graphics which are not available in the text only version. 4.3. PRIDE COURSE 1 (PRIDE-4) The first PRIDE course was held May 31st in Amsterdam. In order to provide room for discussion and useful feedback from the attendees attendance had been restricted to 12. Some of those interested had to be turned down and only one person from each organisation could participate. The major- ity of participants came from service providers, half of those from R&D providers. The general structure of this one day, four session course roughly follows the PRIDE Guide: Morning 1. Basics and Principles of Policy Based Routing 2. Routing Registry Principles and the RIPE RR Afternoon 3. Using the RIPE RR and the PRIDE Tools 4. Future Developments (CIDR, ...) Page 4 May, 1994 PRIDE Report 4 pride-rep4.txt The material for each attendee consists of a slidebook con- taining a copy of the slides with room for notes, a reader containing ripe-81 and its successor in draft form as well as a copy of the PRIDE guide. The material is currently available on request only. We decided to present a fair amount of basic material which should be known to participants at least to some extent. The intention of this is to create a common basis and termi- nology to explain the RR and the tools. After each session and at the conclusion of the course each attendee was asked to complete a questionnaire, an example of which can be found in Appendix G. The average results are presented below: +--------------------------------+--------------------+--------+ | | Sessions | Whole | | +----+----+-----+----+ | | | S1 | S2 | S3 | S4 | Course | +--------------------------------+----+----+-----+----+--------+ |% of material known | 87 | 79 | 42 | 38 | 61 | |% of known material useful | 79 | 81 | 90 | 71 | 88 | |% of new material useful | 97 | 97 | 100 | 89 | 94 | |% of material understandable | 95 | 94 | 98 | 85 | 91 | |satisfaction with slides | 89 | 95 | 97 | 90 | 91 | |satisfaction with presentations | 88 | 83 | 92 | 92 | 93 | |overall quality | 88 | 91 | 91 | 91 | 91 | +--------------------------------+----+----+-----+----+--------+ Table 2: Results of PRIDE course questionnaire From the answers to the "% of known material useful" ques- tion it can be seen, that our approach which repeats some known material is well received. We are quite pleased with the results of the "overall quality" question. An average overall satisfaction of around 90% on the first edition of a course is not a bad result at all. Consequently we will not change the general structure of the PRIDE course version 1. Based on the comments from the attendees we will make some slight adjustments to sessions 2 and 3, introducing more examples and step-by-step walk-throughs into session 3. A second course will be held in July, location and date to be determined by user demand. 5. Revised Work Plan Because of the unexpected work on ripe-81++ and the impact of classless routing on the tools the workplan of PRIDE has been changed. Most noteworthy the project duration will be extended with two months until week 44/94. This has no bud- getary consequences because we still have 4 unused FTE/months from 1993. Because of the delays inherent in the May, 1994 Page 5 PRIDE Report 4 pride-rep4.txt ripe-81 revision process an extension is more appropriate than adding additional resources to the project earlier. The revised workplan is now as follows: +-----------------------------------------------------+ | Milestones Work | +-----------------------------------------------------+ |Week 25 finalise ripe-81++ draft | |Week 27 complete RR software changes for classless | | revised course-1 material | |Week 28 hold another course-1 | |Week 29 beta-release classless tools-1 | |Week 31 ripe-81++ accepted by RIPE routing group | | | |Week 36 release guide-2 | | beta release prconfig | |Week 37 present work at RIPE 19 | |Week 38 tools-2 release (including prconfig) | | | |Week 40 course-2 release | |Week 41 hold courses | |Week 42 hold courses | |Week 43 hold courses | | | |Week 44 final report | +-----------------------------------------------------+ Table 3: Revised work plan 6. Progress Against Milestones PRIDE-2 (PRIDE-TOOLS-1) was completed in week 50/93. PRIDE-3 (PRIDE-GUIDE-1) was completed in week 22. PRIDE-4 (PRIDE-COURSE-1) was completed in week 22. Page 6 May, 1994 PRIDE Report 4 pride-rep4.txt 7. Appendices The list of appendices are a regular part of the monthly reports. They are expected to grow as the project evolves. Appendix A -- PRIDE documents This contains the list of PRIDE documents. The PRIDE project proposal: ftp.ripe.net:pride/docs/pride-prop.{ps,txt} PRIDE Guide (PRIDE-3). ftp.ripe.net:pride/docs/guide-1.0.tar.txt.Z ftp.ripe.net:pride/docs/guide-1.0.tar.ps.Z Appendix B -- Related documents This contains a list of all PRIDE related documents. RIPE Routing Registry format - RIPE 81 ftp.ripe.net:ripe/docs/ripe-docs/ripe-081.ps RIPE-81++ ftp.ripe.net:ripe/drafts/ripe-81++.ps Proposed RR syntax from Merit rrdb.merit.edu:pub/meritrr/policy_syntax.txt Guarded Field Support in the RIPE database - RIPE 108 ftp.ripe.net:ripe/docs/ripe-docs/ripe-108.ps SWIP proposal merit.edu:/pub/nsfnet/swip/swip.txt Appendix C -- PRIDE tools PRIDE-1 (First release of PRIDE-tools 1) ftp.ripe.net:pride/tools/pride-tools-1.tar.Z The pride ftp directory contains the following directories: tools May, 1994 Page 7 PRIDE Report 4 pride-rep4.txt PRIDE tools and related software reports PRIDE monthly project reports docs PRIDE and related documents Appendix D -- Routing Registry (RR) Status State: May 1994 Status of European ASes in RR # of ASes Percentage ----------------------------------------------------------------------------- In RIPE database with Routing Policy information 111 82 % In RIPE database without Routing Policy information 13 10 % Not in RIPE database but in NIC/related databases 12 8 % ----------------------------------------------------------------------------- Total 136 Table 4: Breakdown of known European routed ASes Status of Worldwide ASes in RIPE RR # of ASes Percentage ----------------------------------------------------------------------------- In RIPE database with Routing Policy information 111 29 % Non-Euro ASes in RIPE Database with RP info 2 1 % In RIPE database without Routing Policy information 13 3 % Non-Euro known without Routing Policy information 251 66 % Unknown in any database 5 1 % ----------------------------------------------------------------------------- Total 382 Table 5: Breakdown of known Worldwide routed ASes Appendix E -- Example usage of Interactive menu % telnet info.ripe.net 4711 Trying 192.87.45.1 ... Connected to ns.ripe.net. Escape character is '^]' PRIDE Tools Server 1 - prcheck This tool checks syntax and consistency of AS objects in RIPE-81 format. These objects describe routing policy and jointly form the RIPE Routing Registry. The server version can only check objects already in Page 8 May, 1994 PRIDE Report 4 pride-rep4.txt the database. If you install prcheck locally you can use it to check local objects (before submitting them) as well. You can also find out if your routing policy is consistent with those of your neighbors. 2 - prtraceroute This tool combines the normal traceroute output with routing policy information from the RIPE Routing registry. The output clearly identifies routing problems. The server version can only trace routes from host ns.ripe.net.If you install prtraceroute locally you can also trace from your host. 3 - prpath This tool will display all possible paths according to the routing policy information registered in the RIPE routing registry between a certain source and destination. q - Quit Enter Selection: 2 This server version will display a quite verbose version of the prtraceroute output. The argument to enter is a destination domain name or IP address. The route from ns.ripe.net to the destination will be displayed. Enter destination host []: ns0.ja.net doing prtraceroute -l -v ns0.ja.net traceroute with AS and policy additions [Jan 4 14:02:47 UTC] from AS1104 ns.ripe.net (192.87.45.1) to AS 786 ns0.ja.net (193.63.94.20) 1 AS1104 hef-router.nikhef.nl (192.87.45.80) [I] 2 3 2 ms 2 AS1103 Amsterdam1.router.surfnet.nl (192.16.183.112) [D1] 2 2 2 ms 3 AS1103 Amsterdam2.router.surfnet.nl (145.41.9.130) [I] 5 3 4 ms 4 AS2043 amsterdam4.empb.net (193.172.4.17) [E1] 6 7 5 ms 5 AS2043 london1.empb.net (193.172.4.5) [I] 22 23 27 ms 6 AS 786 int-gw.ulcc.ac.uk (193.172.27.14) [E1] 24 26 33 ms 7 AS 786 ns0.ja.net (193.63.94.20) [I] 45 24 56 ms AS Path followed: 1104 1103 2043 786 AS1104 = NIKHEF-H AS1103 = SURFnet IP AS2043 = European Multiprotocol Backbone AS 786 = The JANET IP Service done. Enter Selection: q May, 1994 Page 9 PRIDE Report 4 pride-rep4.txt Goodbye Connection closed by foreign host. Appendix F -- Example course questionnaire ____________________________________________________________ Questionnaire ____________________________________________________________ Dear PRIDE course attendee, the PRIDE team would like your feedback on the usefulness of the PRIDE course to you. Please take a few minutes to answer the seven questions below. We would like to know who the comments come from as well, so we can maybe discuss some details with you later via electronic mail. This will help us to improve the course. We are measuring the usefulness of the course and not your intelligence, aptitude or level of sleepiness today! Name: Session: Whole course Answers to the following questions are on a percentage scale with five possible answers. Please circle the closest answer. 1. How much of the material presented was already known to you? (100%=I knew everything before, 0%=everything was new) 0% 25% 50% 75% 100% 2. If you consider the material you already knew before; how much of it was useful to you in this course, because it refreshed your memory or was presented in a new way? (100%=all repeated material was useful, 0%=no repeated material was useful) 0% 25% 50% 75% 100% Page 10 May, 1994 PRIDE Report 4 pride-rep4.txt 3. If you consider the material that was new for you; how much of it was useful to you? (100%=all new material was useful, 0%=no new material was useful) 0% 25% 50% 75% 100% 4. If you consider all material presented; how much of it was presented in a way that was understandable to you? (100%=everything was clear, 0%=all was totally obscure) 0% 25% 50% 75% 100% 5. How much were you satisfied with the quality of the slides ? (100%=fully satisfied, can't be done better, 0%=slides were rubbish) 0% 25% 50% 75% 100% 6. How much were you satisfied with the quality of the presentation? (100%=fully satisfied, can't be done better, 0%=unin- telligible presentation) 0% 25% 50% 75% 100% 7. How much were you satisfied whit the overall quality of the course ? 0% 25% 50% 75% 100% Thank you for your help. May, 1994 Page 11 PRIDE Report 4 pride-rep4.txt The PRIDE Team Page 12 May, 1994