THE INTERNATIONAL INTERNET--GLOBAL CONNECTIVITY
After spending several weeks working in Italy this past fall, during which time I had occasion to use the Internet for a variety of purposes, I looked into the status of the international Internet and its future directions. As the "Electronic Seismologist", I thought I would pass along some of my personal experiences, both good and bad, as well as my estimates (guesses) about the future of this part of the Internet.
When away from home, whether just a few states away or across the ocean, travelers with access to the Internet can enjoy many amenities not available a few years ago. For example, while international phone calls are now little more hassle than domestic ones, the combination of their cost and typical time-zone differences makes them less desirable than e-mail messages for many purposes. In my case, before leaving for and after returning from Italy, the Internet provided me with useful e-mail communications with my hosts in Rome, Naples, and at the Istituto Internazionale di Volcanologia in Catania, Sicily. Just before leaving the U.S., I received from Sicily a request to give a survey lecture on "volcano monitoring in the U.S." I was about to e-mail them a return note indicating that I was no expert in U.S. volcano monitoring, other than in the Cascades, and had little material to use. However, I surfed through the volcano Web sites on the Internet using the very good starting point of the Michigan Technological University Volcano Pages (http://www.geo.mtu.edu/volcanoes/) and got a quick overview of much of the monitoring efforts for U.S. volcanos. By making color overheads of a number of the figures I found this way, I put together a general talk which was quite adequate for the purpose.
A surprising additional early use of the Internet was when I ran across an informal advertisement on the VOLCANO listserver mailings for a vacation apartment on Lipari, the largest of the Aeolian Islands in the Tyranian Sea. I contacted the sender of this message and arranged a very inexpensive one-week stay for my wife and myself in a beautiful apartment overlooking a small boat harbor. A computer image of the site was even provided to me ahead of time. The contact person met us at the dock, showed us around the islands, and shared lots of interesting information about the local volcanoes. (To sign up for e-mail from this listserver, send e-mail to firstname.lastname@example.org with no subject and the message "SUBSCRIBE VOLCANO <Your full name>".)
After a meeting in Rome and a vacation in Lipari, I settled in at the Istituto in Catania to get some work done studying the relative locations of volcanic earthquakes at Mount Etna. While using a Sun workstation there, I was able routinely to download my e-mail from my home machine in Seattle and keep up with what was going on back home. CNN is not available in Catania, and English-language newspapers are hard to find, but news of interest was available via Web news servers and special listserver mailings. My wife even obtained routine reports on the amazing success of the Seattle Mariners baseball team during their playoff run. When requested to give a seminar at the University of Catania on Pacific Northwest tectonics, I supplemented the figures I had brought with me with a few up-to-date figures generated by colleagues in Seattle and transferred to me via FTP in PostScript format.
In most respects, using the Internet internationally is no different from using it within the U.S. The primary IP services (telnet, FTP, SMTP, HTTP, etc.) are universal everywhere, and English is the most common language used. However, my recent experiences in Italy and subsequent inquiries indicated that international connectivity speeds are quite variable, are not symmetric, and are usually highly dependent on time of day. Connectivity speeds, as measured by the UNIX program "ping" between Seattle and Catania, varied from about 1/3 second and less than 5% packet loss, to over 1 second and more than 25% packet loss. During busy times (morning on the U.S. west coast) connections initiated from Seattle were uniformly faster and more reliable than those initiated from Catania. In some cases trying to get a 100 KB file from Seattle to Catania would take over half an hour and in others it would take only a few minutes. Most of the delay and periodic bottleneck appear to be in the jump across the Atlantic, though there are certain connections even within Europe that are greatly overloaded.
Just as within the U.S., international portions of the Internet are undergoing rapid and continuous change. As of July 1995 the "core" Internet (with direct IP interconnectivity) was estimated to reach 90 countries and included almost seven million hosts. Peripheral to the core Internet are many other networks, without direct IP connectivity, but which share some gatewayed services such as e-mail. Including these connected-yet-non-IP networks extends the reach of some Internet services to 168 countries and 20-30 million individual users.
Of course, the developing ubiquity of the Internet is not the same everywhere. In many respects, the international Internet works and grows very much like that in the U.S. In some places it's just a year or two behind us, but in others it's about the same or bit ahead of us. Research institutions in the industrialized western Pacific countries (Japan, Australia, New Zealand), as well as in western Europe, generally have Internet access similar to that in the U.S. Latin America and much of Asia have rapidly expanding direct connectivity, while much of Africa and some of Asia still have no (or only limited) direct connectivity, but they do have some services such as e-mail through other networks. The growth of the international Internet, whether measured in terms of number of subnetworks, hosts, or network traffic, is roughly doubling every year. The U.S. expansion rate is currently less than this rate, and some of the newer areas being covered are experiencing a growth rate much higher than this.
Some details from western Europe are interesting. Connections for academic and government research labs have been mostly completed just in the past year or so, but only a few commercial connections and even fewer private ones exist. In the U.S., the commercial and private growth of the Internet has been well underway for over a year. In Europe, support for the operations of the Internet is still a consortium of scientific and different governmental groups, while in the U.S. governmental support is being phased out and commercial support is becoming dominant. In the U.S. relatively inexpensive high-speed, long-distance backbone lines carry data across the country at speeds of 45 Mbps (Megabits per second) via several different commercial carriers with data exchange points in several places. In Europe, while some high-speed lines exist within a few countries, the trans-Europe backbone has a maximum rate of only 8 Mbps. Obtaining reasonably priced trans-Europe higher-speed circuits has been difficult because of the pricing structures of the telephone companies which, in most countries, operate without competition. This is the reason that connections within Europe can be either very good or marginal, depending on the particular combination of links used.
For most seismologists, our international connections are either for communications with colleagues at foreign research institutions or to obtain seismic data from remote seismic networks or stations. For example, the IRIS SPYDER system uses international Internet connections to Europe, Australia, Japan, and New Zealand to obtain waveform data from 29 of the 58 non-North American seismograph stations in its call-up list. Over the past year, the speed and reliability of these Internet connections has declined, until the present when back-up direct phone calls are being placed to collect these data from time to time. The connections across the Atlantic or Pacific will probably continue to degrade in the near future until better international agreements can be reached to share the cost of these expensive segments.
Over the next year or two I expect connections to remote parts of the international Internet will degrade at a higher rate than the degradation of the U.S. Internet. Most intercontinental Internet traffic passes through the U.S., so our international connections end up serving much of the world. In both the domestic and international cases, improvements in the more distant future are likely. As our seismological uses, and indeed research uses in general, become a smaller and smaller part of the total Internet bandwidth, other factors and needs, particularly as the commercialization of the Internet spreads overseas, will drive improvements, which we will realize. In the meantime, other than periodic problems on specific connections and a generally slower data transfer rate, the Internet, both domestic and international, should continue to be a useful tool to us.
Information obtained for this article is from personal experiences and "Computer Networking: Global Infrastructure for the 21st Century" by Vinton G. Cerf, "Bottom-Up Information Infrastructure and the Internet" by Anthony-Michael Rutkowski, and "Towards a European High-Speed Backbone" by Michael Behringer. All papers were presented at the INET'95 Conference, Honolulu, Hawaii, June 27-30, 1995 (http://www.isoc.org/HMP/index.html).
SRL encourages guest columnists to contribute to the "Electronic Seismologist." Please contact Steve Malone with your ideas. His e-mail address is email@example.com.
Posted: 10 February 1999