Bert van Pinxteren CCIRN Meeting, 26 April The TERENA NREN Compendium, 2000 – 2006: making the case for NRENs (plus some candid comments) Bert van Pinxteren TERENA
Bert van Pinxteren CCIRN Meeting, 26 April But first… TERENA has more activities that help to make the case for NRENs, for example the assistance to the development of research and education networking in less-advanced regions that is part of the GN2 project Country needs assessments carried out for/with Albania, Morocco, Romania, next one will be Moldova More information at support/ support/
Bert van Pinxteren CCIRN Meeting, 26 April Overview History/approach Difficulties Results How you can help
Bert van Pinxteren CCIRN Meeting, 26 April History Trial edition published in limited number of copies by TERENA in 2000; Two parts: country-by-country info only on web, compilation of data on web and in print First real edition published in 2001 (printed part: 24 pages, 30 NRENs); Consistent support and encouragement from the European Commission; 2005 edition part of the GN2 project (printed part:106 pages, 49 NRENs); Website with 2000 – 2006 data:
Bert van Pinxteren CCIRN Meeting, 26 April Basic approach: Organisation-oriented rather than connection-oriented. (A more connection-oriented approach is the ARENA project, see
Bert van Pinxteren CCIRN Meeting, 26 April Role of the Compendium The fast train itself: the networks / GÉANT Preparing for the future: SERENATE and EARNEST Monitoring the progress: the Compendium
Bert van Pinxteren CCIRN Meeting, 26 April Procedure for producing the Compendium (1) Review panel – as diverse as possible (different roles, different cultural/linguistic backgrounds) Current composition: –Marko Bonač (ARNES, Slovenia) –Sabine Jaume-Rajaonia (RENATER, France) –Mike Norris (HEAnet, Ireland); –Esther Robles (RedIRIS, Spain); –Lars Skogan (UNINETT, Norway).
Bert van Pinxteren CCIRN Meeting, 26 April Procedure for producing the Compendium (2) Compile feedback received from last time Discuss in Review Panel, develop questionnaire in close consultation with Review Panel Consult GN2 activity leaders, other who are interested Develop web interface, import data from last year, ask Review Panel members to test the interface Announce to NRENs, remind people ceaselessly Publish data received on the web Review data received and ask additional questions if data seem odd (for example: Terabytes and Terabits) Ask NRENs themselves to review their data Used to prepare draft compilation of data for TERENA members only, presented at the GA meeting (this has been abandoned since 2005)
Bert van Pinxteren CCIRN Meeting, 26 April Procedure for producing the Compendium (3) Get feedback and corrections from NRENs; Get feedback and corrections from Review Panel and TERENA technical staff; Ask for clarifications where needed; Produce final version of part 2. People involved: –Project leader (speaking) –Database designer / webmaster (Jeroen Houben) –Layout person (Carol de Groot) –Content support (TERENA technical staff) –From 1 April: Data Analyst (Maarten de Jong)
Bert van Pinxteren CCIRN Meeting, 26 April Procedure for preparing the Compendium (4) Not wholly uncontroversial: From the minutes of the GA in May 2001: Jan Gruntorád felt that the document contained interesting data, but also contained some errors. He regretted that it had been presented at the conference before the NRENs had had a chance to review the document. From the minutes of the GA in June 2002: Several speakers pointed to tables that could easily lead to misinterpretation, if seen in isolation. Bert van Pinxteren clarified that the tables should be seen as starting points for a discussion.
Bert van Pinxteren CCIRN Meeting, 26 April Procedure for producing the Compendium (5) Advantages of this procedure: –Ensures good response rate –Annual publication helps – it is important to be included –TERENA is trusted –Questionnaire gets better every year –We understand the NREN community and can therefore safeguard against hasty comparisons that do not show the full picture –NRENs get something in return that they can use nationally –Actually reduces questionnaire overload –Becoming a repository of historical data
Bert van Pinxteren CCIRN Meeting, 26 April Difficulties Differences in interpretation and in national situations, e.g.: –What is a university? What is a university site? –How to compare budgets? –Who pays for which level of networking? –What to count and what not to count?
Bert van Pinxteren CCIRN Meeting, 26 April Final eEurope Benchmarking Report, COM(2002) 62 final of February 2002:
Bert van Pinxteren CCIRN Meeting, 26 April Topology of the RedIRIS Network, 2001/2002
Bert van Pinxteren CCIRN Meeting, 26 April Comparison perhaps not completely fair… RedIRIS had 17 connections at 155 Mb/s GARR had just upgraded 3 centres to 2.5 Gb, but each via 2 routes, so they claimed to have 5 Gb/s Many other centres in Italy in fact connected only at 34 Mb/s (yellow lines)
Bert van Pinxteren CCIRN Meeting, 26 April Difficulties (2) Getting (almost) all NRENs to respond requires considerable persuasive skills Finding the balance: more questions or more answers? Finding the balance: better questions or better comparability with earlier years? It does take a bit of work
Bert van Pinxteren CCIRN Meeting, 26 April Results (1) Dany Vandromme, RENATER: As example, I would mention the Compendium (…), which turned [out] to be extremely useful to RENATER, to provide my national authorities with (…)
Bert van Pinxteren CCIRN Meeting, 26 April Results (2) Single place where information can be found Information from earlier years also there Interest from policy makers (Commission), researchers, companies
Bert van Pinxteren CCIRN Meeting, 26 April Results (3) Presented at several national NREN Conferences, NORDUnet and TERENA Conferences Used by NRENs for a variety of purposes: for rectors, funding bodies, benchmarking studies (UKERNA) etc.
Bert van Pinxteren CCIRN Meeting, 26 April Results (4) In the beginning, we were very careful with formulating conclusions from the materials. A few attempts were presented at Conferences, but people almost always were able to point to weaknesses in the analysis Gradually, we have become better at seeing what can be concluded on the basis of the material; most of that is obvious to the NRENs, but not to the outside world EU has wanted us to do this
Bert van Pinxteren CCIRN Meeting, 26 April Summary of key findings: Legal form Users/clients Network Traffic Services Funding
Bert van Pinxteren CCIRN Meeting, 26 April Users/clients Gigabit Ethernet is being introduced by many hitherto less developed NRENs (such as AMREJ, MARNET and RENAM) and thus seems to make it possible, for the first time, to quickly address an important aspect of (…) the digital divide in Europe (…).
Bert van Pinxteren CCIRN Meeting, 26 April Network There are indications that more and more NRENs are switching over to dark fibre as the technology of the future, with the EU NRENs being in the lead The trend seems to be that in the more advanced countries, the core capacity will evolve to 10 Gb/s or multiples of that.
Bert van Pinxteren CCIRN Meeting, 26 April Traffic (1) Growth rates in the new EU member states and in non- EU/EFTA countries are clearly higher than those in the old EU member states It seems that in the EU, traffic growth is slowing down
Bert van Pinxteren CCIRN Meeting, 26 April Traffic (2) It seems that traffic is now determined more by (changes in) user demand, rather than by network capacity limitations. It is unclear if this trend towards slower growth will persist The longer-established NRENs from the EU and EFTA countries are mostly net exporters of data, while the other NRENs are net importers.
Bert van Pinxteren CCIRN Meeting, 26 April Traffic (3): congestion index
Bert van Pinxteren CCIRN Meeting, 26 April Funding Comparisons are difficult (almost impossible) The trend is that budgets stay relatively stable and that NRENs are able to deliver more bandwidth and more services for roughly the same amount of money. EU projects seem to act as a catalyst for increased national NREN budgets in some countries
Bert van Pinxteren CCIRN Meeting, 26 April Bonus material /login.phphttps:// 2005/login.php
Bert van Pinxteren CCIRN Meeting, 26 April Bonus material Compendium spreadsheets home :: compendium :: 2005 :: spreadsheets These spreadsheets are posted here as an experimental service only for the NRENs that have participated in the survey. They can be used for example for preparing national reports that contain data from a subset of NRENs that is relevant for your national situation. Recognising TERENA as the source of the data will be appreciated. Note that most of the data have been collected in March and April of If you find new, nice ways of presenting the data, please feel free to share your insights with us. It may take some studying to figure out what is what in the spreadsheets - but if you have trouble with them, you can always ask for clarification. All spreadsheets have a column with the name of the NREN and the country. Most of them also have some way of sorting the countries geographically, as in the Compendium. E.g. and 'a' might be used for EU countries, a 'b' for other countries. Note that these distinctions have not been used consistently throughout the spreadsheets.homecompendium2005spreadsheetsshareask Basic information –Legal form of NRENs, partner organisations _6.20_organisation_info.xls6.19_6.20_organisation_info.xls Users / clients –Connection policies, percentages of institutions in different categories served by the NREN - Ch._2.8_inst_connect_marketshare_graphs.xls Ch._2.8_inst_connect_marketshare_graphs.xls –Uptake of IPv6 - ipv6_unis_research_1.5_1.7.xlsipv6_unis_research_1.5_1.7.xls –Numbers of connected institutions and bandwidth - 1.5_inst_conn_speeds.xls1.5_inst_conn_speeds.xls Network –Core capacity - Ch.3.2_Core Capacity00-07l.xlsCh.3.2_Core Capacity00-07l.xls –Core network size - Ch.3.4_Core Network Size multiyear.xlsCh.3.4_Core Network Size multiyear.xls –External links - Ch.3.5_external_links.xlsCh.3.5_external_links.xls –Dark fibre - Ch.3.7_darkfibre.xlsCh.3.7_darkfibre.xls Traffic –External traffic ('T3' and 'T4')- Ch.4.2_traffic_t3_t4_multiyear.xlsCh.4.2_traffic_t3_t4_multiyear.xls –Traffic load - Ch.4_links_load_for_final.xlsCh.4_links_load_for_final.xls –Congestion - 3.5_congestion.xls3.5_congestion.xls Services - 5.3_5.4_5.6_5.7_services.xls5.3_5.4_5.6_5.7_services.xls Tasks, Staffing, Funding - 4_funding_staffing.xls4_funding_staffing.xls Appendix –Number of PoPs and of managed links on the network - 2.1_2.2_network.xls2.1_2.2_network.xls –Traffic with the general Internet - external_traffic_3.3_3.4.xls (use columns D and E)external_traffic_3.3_3.4.xls
Bert van Pinxteren CCIRN Meeting, 26 April How YOU can help (1) We would like to get some information also from NRENs in other parts of the world, to provide a more complete picture for ourselves (and maybe it is also useful for you) For this, we have developed a mini-version of the survey, containing a number of questions taken from the main survey, with focus on: connection policies and market share; basic network characteristics; traffic; funding; basic contact info.
Bert van Pinxteren CCIRN Meeting, 26 April How YOU can help (2) We would like as many NRENs as possible to enter data in the mini-survey, but NRENs are free to complete the full questionnaire if they want to. Procedure: let us know you are willing to participate -> we will make a record for your NREN in the database Request a password at Complete the survey, double-check your answers! More information at fo.html fo.html
Bert van Pinxteren CCIRN Meeting, 26 April What you will get in return The printed version of the Compendium Help in presenting the data Other benefits as discussed earlier Of course, we are open to working with others who want to set up similar but adapted initiatives!