From the Editor-in-Chief . . . The Intelligence Within Munindar P. Singh • [email protected]

BEING INTERACTIVE The Internet provides the backdrop Antithesis for an important technical controver- The antithesis is the vision of the stu- sy in the domain of telecommunica- pid network. This vision is promulgat- tions. Like all good technical contro- ed by Internet anarchists and free versies, this one is political as well. enterprisers alike, who would like to The controversy is interesting to see freedom blossom, specifically free- computer scientists because it helps dom to introduce new services. focus our thinking on system design George Gilder,1 David Isenberg,2 and and evolution, and because it gives us others have long argued that as band- an opportunity to reflect on the width becomes plentiful, intelligence nature of the Internet. To highlight will propagate to the edges of the net- some differences in a brief format, I work and the network itself will pro- will present the controversy in the vide no more than bit transport. style of the Hegelian dialectic (admittedly, though, The overwhelming advantage of stupid networks is with oversimplifications). that, like the Internet, they naturally support hetero- geneity and extensibility. End users can choose which- Thesis ever applications they like and invoke whichever ser- The thesis is the vision of the intelligent network as vices they like without requiring consistent changes promulgated by the traditional telephone operators: throughout a large network. You can plug in a new the telecom system as a big beast, central to every- service as easily as you can plug in a Web server. thing you do that pertains to communications. If you don’t like it, the thesis says, we (and only we, Analysis (OK, so maybe I should credit Freud as well as Hegel.) There is, as they say, truth on both sides. The first argument in favor of the thesis is efficiency. A managed network can deliver better performance The real question isn’t than an unmanaged one. But the disruption offered about intelligence by communications technology plays an important role here. I am alluding to the rapid improvements but about innovation— in bandwidth supported by optical technology. For on the edges and deep within. the past few years, these improvements have exceed- ed those in computing power.3 There is more band- width than we can hope to spend, so letting some of it go to waste in exchange for simplicity and open- ness is a good trade-off to make. That is, the effi- not you) can make it better, which we will do by ciency argument for the thesis doesn’t stand. embedding more and more intelligence in it. As the On the other hand, the traditional telecom beast becomes smarter, it can offer newer “services,” approach may be difficult to manage, run, and which look, from the consumer standpoint, like fea- evolve, but at least it can offer guaranteed reliability. tures—caller ID, call forwarding, and so on. The telecoms are legendary for their “six nines” Telecommunication companies have been remark- availability. The Internet, by contrast, is what we— ably successful in carrying this vision forward. Not rather euphemistically—call a best-effort system. only is the network more intelligent, it is also highly Internet services, in general, make no guarantees, reliable and appears to scale well. Unfortunately, this not even of “six ones” availability. approach also makes it harder to add services. To function properly, a new service requires changes in Synthesis several places internal to a large (and apparently So should we pick one or the other doctrine? Far monolithic) system. from a straightforward dichotomy, what we see is

6 NOVEMBER • DECEMBER 2000 http://computer.org/internet/ IEEE INTERNET COMPUTING IEEE INTERNET COMPUTING IEEE Computer Society Publications Office 10662 Los Vaqueros Circle, PO Box 3014 Los Alamitos, CA 90720-1314 each side influencing the other. Telecom networks are becoming more and more like the Internet, not only in EDITOR-IN-CHIEF being oriented toward packets instead of circuits, but Munindar P. Singh • [email protected] also in becoming open with their services and interop- ASSOCIATE EDITOR-IN-CHIEF erable with other telecom networks. These changes are Robert Filman • rfi[email protected] most prominent in the arena of mobile third-genera- EDITORIAL BOARD tion networks (for more information, see Salah Aidarous • [email protected] http://www.umts-forum.org/what_is_umts.html). (IEEE Communications Society Liaison) But intelligence isn’t just moving outward; some of it Donna Auguste • [email protected] is also moving inward. Over the past few years, private Miroslav Benda • [email protected] networks—physical or virtual—have become a norm over the Internet. Only in these private networks can Elisa Bertino • [email protected] we hope to make any guarantees for performance or Scott Bradner • [email protected] quality of service. We can’t guarantee compliance with a Fred Douglis • [email protected] service level agreement unless we provision and manage Stuart I. Feldman • sif@us..com the network. These networks are more than bit trans- Li Gong • [email protected] ports; they can act intelligently and also support impor- Seif Haridi • [email protected] tant services such as billing and micropayments. Michael N. Huhns • [email protected] At the same time, there are services the traditional Ray W. Johnson • [email protected] networks can support more naturally, for example, the Gail E. Kaiser • [email protected] location of a mobile device. So it is appropriate to pro- Leonard Kleinrock • [email protected] vision these services from the inside, essentially by Doug Lea • [email protected] moving the intelligence inward. Frank Maurer • [email protected] The real question, of course, isn’t about intelligence but about innovation. The Internet succeeds because it Charles E. Perkins • [email protected] Charles J. Petrie • [email protected] supports—even forces—innovation, not only on the (EIC emeritus) edges of the network but sometimes deep within. Agostino Poggi • [email protected] References William Regli • [email protected] Helmuth Ritzer • [email protected] 1. G. Gilder, “George Gilder on the Bandwidth of Plenty,” IEEE Internet Computing, vol. 1, no. 1, Jan. 1997, pp. 9-18. Anthony Michael Rutkowski • [email protected] 2. D.S. Isenberg, “The Dawn of the Stupid Network,” ACM Ravi Sandhu • [email protected] Networker, vol. 2, no. 1, Feb. 1998, pp. 24-31. Brian Thomas • [email protected] 3. L.G. Robertson, “Packet Switching or Optical Switching,” STAFF IEEE Internet Computing, vol. 4, no. 1, Jan. 2000, pp. 50-51. Managing Editor: Linda World [email protected] Associate Editor: Joan Taylor [email protected] IC Adds Two New Board Members Staff Editor/IC Online Design: Steve Woods Donna Auguste is president and CEO of Freshwater [email protected] Software, a company that develops Web manage- Magazine Assistant: Hazel Kosky [email protected] ment tools. She has a BA in electrical engineering Art Director: Joseph Daigle and from the University of , Berkeley, and has completed graduate Cover Design: Joseph Daigle studies in computer science at Carnegie Mellon Graphic Artist: Ken Duckworth University. She currently leads an effort to bring Contributing Editors: David Clark, Keri Schreiner, Martin Zacks solar-powered electricity and e-mail to hospital clin- Publisher: Angela Burgess ics in remote villages in Africa. Membership/Circulation Marketing Manager: Georgann Carter Elisa Bertino is professor of database systems in the Business Development Manager: Sandy Aijala Computer Science Department of the University of Advertising Supervisor: Marian Anderson Milan where she is currently department chair. She CS MAGAZINE OPERATIONS COMMITTEE has published more than 200 papers on her research in database sys