On the Euro-City train to Paris, I started to feel a little better. The conductor looked at my Eurail pass and smiled broadly.
“Ah, un American! C’est tres bon!” he exclaimed with a classic Parisian accent. This was about as out of character for a Parisian as it would be to have a New Yorker tell you to “have a nice day.” Still, it was a welcome diversion from the dour gloom of Belgium.
As the train passed through St. Quentin and headed up the Oise river, the countryside started getting familiar, the houses looked French, and even the farms had a characteristic look. I arrived at the Gare du Nord and stepped into the car of a virtuoso taxi driver who shot down alleys and back streets, delivering me to the Montparnasse district for the remarkable fare of FF 70. I gave him 100, the amount such a ride would normally cost.
Flipping on CNN in my room, I was just in time for a commercial featuring “beautiful, bountiful, beguiling, Belgium.” Popping open a beer from the minibar, I mentally added “bullshit and bureaucrats” to their list.
After watching President Bush’s State of the Union address, I bundled up and went up the Boulevard Montparnasse to fetch my 92-year old great-aunt, stopping on the way for a paper sack of roasted chestnuts to warm my hands and popping into Le Chien Qui Fume for a quick drink.
Over a salad of preserved quail on a bed of dandelions, watercress, onions, and gooseberries, I tried to explain to my aunt the concept of a technical travelogue. She listened carefully, frowned a bit perplexedly, and told me that she was very proud of me, whatever it was that I was doing.
Thursday morning, I took the Métro to the edge of the city, then a train out to Versailles. My taxi took a detour around the immense castle that had housed Marie Antoinette and over to the neighboring suburb of Rocquencourt, where INRIA has a facility.
My 10 A.M. appointment wasn’t there. He had sent me mail, but when I logged into a borrowed terminal I saw, to my horror, that I had received no new mail messages for over a week. Something was drastically wrong.
The secretary at INRIA told me where I had to go for my meeting. On the way back into town, I tried to figure out what had happened. The problem was no doubt in my home PC (or, more likely, in the way I had configured it).
The domain of Malamud.COM is a registered domain. When somebody starts to send mail to any of the millions of possible addresses in the Malamud.COM domain, a record in the Domain Name System points to my commercial service provider, Colorado SuperNet.
Colorado SuperNet takes all incoming mail and spools it on a special UUCP-only account I maintain on their machine. A few times a day, my PC calls up that account and retrieves all incoming mail messages. Those messages are sorted, combined with my MCI Mail, any fax notifications I can grep out of my fax log, and some miscellaneous status messages, and the whole lot is sent right back up to the interactive account on Colorado SuperNet that I use while on the road.
In desperation, I stopped at the train station and gave a call to my fax number. Since the fax is answered by a fax board in a PC, it stands to reason that a high pitched tone on the other end would tell me that my PC was still up and running, and, by logical conclusion, that my house was not a charred, smoking wreck.
Having successfully pinged my house (but not solved the mail problem), I continued on to La Défense, a complex of convention centers, shopping centers, exhibit halls, and other public spaces, all surrounding a futuristic looking arch, kind of a 21st-century rendition of the Arc de Triomphe. There, next to the World Trade Center, is the InfoMart, a three story high set of exhibit spaces for computer vendors.
Most of the vendors are your typical IBM, Bull, and Microsoft variety, but there are a couple of odd ducks. One is a “house of the future” exhibit, built to the new European home automation standards. This house-like display, full of spas, jacuzzis, exercise bikes, and other things I would never allow in my own house, was a cooperative venture of a dozen vendors, all trumpeting the future of domotics.
The other shop that doesn’t really fit in is one rented by four research agencies, including INRIA. There, I met with Milan Sterba, a young Czech with joint appointments at INRIA and the Prague School of Economics.
Though officially he had been in the business school in Prague, Milan had spent most of his time on the large Czechoslovakian project to develop an MVS-like operating system to run on the reverse-engineered System 370 clones manufactured in Bulgaria and Russia. Milan had spent several years working in his chosen specialty, telecommunications, developing a VTAM-like telecommunications access method.
At INRIA, in addition to real work like reconfiguring Sun workstations, Milan had an informal role as one of the focal points for East European countries trying to get on the Internet. He maintained a document detailing current connectivity, chaired sessions at RIPE meetings, and otherwise helped to spread information around where it was needed.
The rapid progress in the former Eastern Bloc and Soviet Union had been truly amazing. As fast as the countries could persuade the U.S. to process the paperwork for Cisco routers, countries were plopping in TCP/IP nodes, enhancing EARN connections, and using UUCP and EUnet to spread connectivity into new places.
Bulgaria, for example, was using a dial-up WCP link to Amsterdam for connectivity. Dial-up was then being used to send messages and files to the 10 Bulgarian EUnet sites. Dial-up WCP was a common first step for many countries, a fact worth keeping in mind before sending long messages to people on the other end of slow (and expensive) lines.
Poland was typical of countries with higher levels of connectivity. Poland had a 9,600 bps leased line between Warsaw and Copenhagen and used statistical multiplexors to combine TCP/IP with the EARN NJE/BSC protocols. Another 9,600 bps leased line between Krakow and CERN used DECnet protocols as part of HEPnet, but was converted to IP as soon as an export license for a Cisco router was approved. A 64 kbps leased line between Warsaw and NORDUnet in Stockholm was going operational in 1992 and would greatly enhance TCP/IP connectivity to the country.
Milan and I talked about the effect that posting standards had had in Czechoslovakia. Milan told me that, to his knowledge, only one copy of the ITU Blue Book existed and people spent considerable time going to a central facility to consult standards documents.
The ITU and others in the standards cartel had always insisted that the standards were reasonably priced for those “serious” about doing work. Milan confirmed that this position was nonsense and that many countries in Eastern Europe had particularly welcomed the ability to access documents they needed for their work.
I found it particularly distressing that the ITU policies were having the effect of preventing people in developing countries from accessing technical standards. After all, one of the purposes of the United Nations, of which the ITU is a key part, is to promote a world community. Keeping key documents hidden from those without money, indeed keeping documents hidden from entire countries without money, is certainly a convoluted perversion of the UN mission.
That night, I had dinner in a Venetian restaurant, a classic place with a dozen tables and the owner acting as the head (and only) waiter. In his 50s, stout, distinguished looking and impeccably dressed, the owner showed people to their tables, practicing the art of being apparently servile while in reality insulting everyone he could. While ignoring my waiter and eating a mediocre mezzaluna funghi, I read Don Fernando, Somerset Maugham’s classic essay on Spain. Maugham was writing about the playwright Lope De Vega, author of 2,200 plays.
Many had proudly pointed to the size of this document base, equating De Vega’s proclivity for producing paper with greatness. De Vega himself had not taken this seriously, remarking that “if anyone should cavil at my plays and think that I wrote them for fame, undeceive him and tell that I wrote them for money.”
When you reward people for producing paper, they will do so. When the European Commission and RARE paid people’s expenses to go to meetings and make specifications, it is not surprising that they took so long to make them, nor that they were so voluminous.
Likewise, when you take a stack of OSI documents and put them next to the RFC series, you can tell pretty quickly which ones were produced by standards professionals and which ones were written by engineers who had software to write and networks to run.
Friday morning, I met Jean-Paul Le Guigner of the Comité Réseau des Universités, a committee of French universities formed to set up a national research network, much like EDUCOM in the U.S. and the vice-chancellors committee in Australia.
We took the subway to Jussieu, changing trains and walking through the long, intricate tunnels connecting Paris Métro lines, ending up at the Université Pierre et Marie Curie. Widely acknowledged as one of the uglier campuses ever built, the university was constructed during the period in the 1960s and 1970s when slabs of unfinished concrete were considered “modern.”
At the university, we found the office of Christian Michau of the Unité Réseaux du CNRS, Jean-Paul’s equivalent in the research world. CNRS, the Centre National de la Recherche Scientifique, acted as a coordinating body for French research institutions like INRIA and ORSTOM.
The two groups, the universities and the researchers, had banded together to form Renater, the Réseau National de Télécommunications de la Recherche. Funded to the tune of FF 50 to 60 million (around U.S. $10 million) per year, this network would form a national backbone of 2 Mbps in 1992 with the core expanding to 34 Mbps in early 1993.
Renater, like the NSFNET in the U.S., was set up as a network of networks, linking regionals together and France to the rest of the world. Unlike NSFNET, the interface to the network would be provided by the telephone company, in this case France Telecom.
France Telecom, under contract to Renater, would offer a point of presence to which regionals could connect. The regionals, like the R3T2 network I had seen in Sophia-Antipolis, were locally funded. In the case of R3T2, for example, local and regional networks had started the network as a way of attracting industry and promoting education.
After checking to see if my house was spitting up mail yet, Christian, Jean-Paul, and I headed over to Christian Michau’s office and sat around a conference table. We were joined by two other gentlemen who were somehow joining Renater and needed a briefing on the project.
It was quickly established that having me speak French was a better strategy than forcing the other four to speak English. Christian suggested I start with a brief description of my project. Somehow, while I was researching this book I found that many of my hosts, though always very hospitable and full of information, never quite understood the concept of a “technical travelogue.”
I complied, giving a little speech and feeling quite content that I had remembered enough French to pull this off. Everybody smiled and nodded gravely, each in turn thanking me. Then, one of the observers leaned over to Jean-Paul and suggested that perhaps a good way to proceed would be for somebody to explain exactly what the American gentleman was hoping to accomplish.
Once this was all straightened out, Christian and Jean-Paul proceeded to tell me about Renater. In France, research institutions and universities have a great deal of independence. To make things difficult, any attempt to establish a national research network would cut across three different government ministries.
Through a minor miracle, in February 1991 an interministerial convention was signed which established the authorization for Renater to begin planning. A pilot committee was formed, which spawned a technical committee, which created a project team. By the end of January 1992, a contract with France Telecom was days away from signing and a pilot network was already in place. Remarkably quick, given the number of ministries and bureaucracies involved.
Renater was designed to be a multiprotocol network, but with an interesting twist. The networks in many countries, such as Germany’s DFN, ran all protocols over X.25 (“the pathway to OSI”). Running TCP/IP over X.25 was certainly a combination that worked, but for connecting two hosts together a straight leased line made more sense.
Renater was set up in a very pragmatic fashion as two networks, one presenting an IP interface, the other an X.25 interface. At 2 Mbps, the X.25 network would simply layer on top of the Transpac public X.25 network. The IP network, a totally separate system, would be a series of Cisco routers on leased lines.
The network would be quickly upgraded to 34 Mbps in early 1993, at least on core routes. It was hoped that at these higher speeds that ATM-based cell switching would allow the X.25 and IP networks to coexist on the underlying substrate.
Presenting an X.25 interface solved several problems. OSI could run on the network, so all parties in Renater were able to boldly proclaim that theirs was an OSI network which incidentally happened to support “immediate needs of existing traffic.” The X.25 network also gave a platform on which DECnet and SNA traffic could run.
Michau and LeGuigner were decidedly pragmatic about their work, trying hard to avoid political battles and concentrate instead on getting the network up and running. They strenuously avoided religious decisions they didn’t have to make, using mechanisms like study groups to look at questions of SMTP versus X.400.
What I found most interesting was the close working relationship with France Telecom. The telephone company seemed to view Renater as an opportunity instead of a threat, using the research community as a place to test technology and building expertise for future commercial rollouts.
A 5-year contract worth roughly U.S. $10 million per year had been signed with the telephone company, giving them a substantial incentive to provide a working network. In addition, a contract provision was added that any tariff decreases in the future would be used to upgrade the bandwidth of the research network, thus guaranteeing a fixed amount for the telco and at the same time protecting the research community.
How had Renater managed to avoid the need to make OSI their only protocol? Michau had one of the better answers to the question of OSI had seen.
“Many countries in Europe are beginning to realize that OSI is not X.25,” he explained. Instead, he looked to standards like OSI to provide solutions to a range of problems, particularly at the application layer.
If you accept OSI as a set of services instead of some monolithic religion, useful protocols like X.400 and X.500 can be applied immediately instead of waiting for an entire grand design to come into being. In fact, you could, with this view, run X.400 on top of ISODE, which in turn can run on top of TCP/IP, and still provide OSI service.
Defining OSI as a series of applications took Renater out of the debate. After all, Renater was simply providing an infrastructure, a core backbone, and it was up to the users to decide what to run on that backbone.
Running an X.25 interface kept Renater in the OSI game. Even for international links, Renater was willing to use X.25 links. While acknowledging that “some choices are not technically perfect,” it was clear that Renater would consider projects like the 2 Mbps IXI extension. After all, the European Commission favored X.25 and the Commission had lots of money. It was clear that the larger European countries like France would quickly catch up to the U.S., putting in high-bandwidth backbones, setting up regional networks, and running multiprotocol environments.
That night, I picked up my great-aunt for a quick dinner. Over glasses of port and a dish of snails soaked in garlic and butter, we talked about her childhood in Russia, her emigration to Montreal and New York in the early part of the century, and her escape from the Nazis in the war, walking over the border to Free France with her 6 year old daughter in the middle of the night. She had been in France for 60 years, even receiving the Legion of Honor for her work at the Pasteur Institute, where she conducted research and later helped numerous visiting scientists get settled. After walking her back to her apartment, I went to my seedy hotel to pack and trade insults with the snooty night manager.