Steve Goldstein: First of all, I’ve been retired since 2003. So, most of the active role I’ve played has been between 1989 and 2003 at the National Science Foundation, with a slight adden­dum because I served on the board of the Internet Corporation for Assigned Names and Numbers—ICANN—from 2006 to 2009. And I keep look­ing at the cam­era instead of you but that’s habit. 

So at the National Science Foundation, my job by…the ele­va­tor speech? My job was to con­nect the Internet, or bring Internet around the world. To expand on that, when I came to the National Science Foundation in 1989, the NSFNET had estab­lished itself pret­ty much as the cen­ter of grav­i­ty of the Internet. And some peo­ple want­ed to say was the core of the Internet, but that’s a tech­ni­cal term. And it real­ly was­n’t because the Internet was a net­work of net­works, so it was no longer a core. But it was a cen­ter of grav­i­ty. Everybody want­ed to con­nect to the National Science Foundation net­work and my job, or my mis­sion that I assumed for myself, was to help them do that. 

And so lit­tle by lit­tle we con­nect­ed a lot of coun­tries, about twenty-five coun­tries between 1991 and 95, with the help of a project that I set up called the International Connections Manager, which was award­ed to Sprint. And Sprint had no expe­ri­ence of the Internet at that time. This was their way of jump­ing in. And they real­ly bent over back­wards to do every­thing pos­si­ble to make things hap­pen, so they were a great part­ner. But we connected—or, well, we solid­i­fied exist­ing con­nec­tions with the Nordic countries—NORDUnet—and with France. That was the begin­ning of things. 

But along the line we helped to con­nect Russia and China. And the very last con­nec­tion before the project was end­ed was with Mongolia. And that was 95, 96. And Latin American coun­tries along the way, with the help of the Organization of American States; Saul Hahn my part­ner in crime there. Saul worked in-country with the var­i­ous gov­ern­ments and was able to pro­vide some mon­ey for equip­ment while on the oth­er hand we worked to help get the satel­lite links con­nect­ed and get those satel­lite links con­nect­ed to the NSFNET with the help of Sprint. So that was a big partnership. 

Then Al Gore came along and start­ed a G7 ini­tia­tive called Global Interoperability of Broadband Networks—GIBN. And the idea there was that broad­band networks—and by that he meant 45 megabits per sec­ond, which today is not fast at all. But they were begin­ning to spring up and he want­ed a way of inter­con­nect­ing those net­work from oth­er coun­tries. And so what we did in response to that was to kick off a project at University of Illinois, Chicago with Tom DeFanti and Maxine Brown to set up a router in Chicago where coun­tries could haul their links across the US and con­nect to each oth­er there. 

Now I men­tion haul­ing across the US. If we had set this thing up in Los Angeles, that would’ve been great for the Asians. Taiwan and you know, Japan and so forth. But then the peo­ple from Europe would­n’t be too hap­py cause they’d have to come clear across the US. If we set up in New York it would have been the same thing with the peo­ple from Asia not being too hap­py. So we set it up in Chicago where nobody was hap­py, or every­body was equal­ly unhap­py. But it worked rather nice­ly and many coun­tries con­nect­ed to the STAR TAP.

And the beau­ty of the STAR TAP was every­body that con­nect­ed to a router. So for exam­ple, France if they want to con­nect with Taiwan through that router with­out going through any inter­ven­ing US infra­struc­ture. So there was no US pol­i­cy imposed. And that was one of the key fea­tures. And this went on to be repli­cat­ed at inter­con­nec­tion points all over the world. So it was one of the first ones there. And that was real­ized in 1997

And then, toward the lat­ter part of the 1990s, Greg Cole at University of Tennessee got a coop­er­a­tion going among Russia, China, and the United States to basically—the con­cept was to set up a ded­i­cat­ed wave­length around the Northern Hemisphere to sup­port computationally-intensive research. And this was called GLORIAD. And GLORIAD is still run­ning today, although many oth­er coun­tries or net­works have joined. So for exam­ple Canada, the Nordic countries—NORDUnet, SURFnet in Holland, Singapore, India, Egypt—the most recent one. And I’m pret­ty sure I’m leav­ing somebody—oh, Canada. And they’re all mem­bers of this GLORIAD group. So GLORIAD’s archi­tec­ture’s now sort of a ring of rings around the Northern Hemisphere. But again, to sup­port computationally-intensive sci­ence like weath­er research and so forth. So that’s a pro­gres­sion that sort of helped. And I was more an advi­sor to GLORIAD toward the end, and still do so infor­mal­ly at times. 

One oth­er thing that’s impor­tant was the grass­roots net­works. Places in the under­bel­ly of Asia, for exam­ple, or Africa, and even some in Latin America, which were just try­ing to start up. And Randy Bush at the Network Startup Resource Center was our awardee for that. Later taken—when Randy went on to oth­er things, Steve Huter, now at University of Oregon, runs the Network Startup Resource Center. And they do a lot of train­ing of net­work oper­a­tors. They lever­age funds from us and now from Google to get books and oth­er mate­ri­als to start­up net­works around the world. 

And even a lot of used equip­ment. So for exam­ple, at some point when peo­ple would trade in equip­ment to Cisco for big­ger and more pow­er­ful stuff, the rumor was—and I don’t know how true this is–but the rumor was that Cisco actu­al­ly bull­dozed this stuff so that it would­n’t be used on the mar­ket because they want­ed to then sell the more advanced. Well, stuff was saved from the bull­doz­er and actu­al­ly shipped to a lot of these start­up net­works in dif­fer­ent parts of the world. Again, the under­bel­ly of Asia and Africa. And Steve Huter at the Network Startup Resource Center was respon­si­ble for that. So, Steve con­tin­ues to get funds from National Science Foundation, and as I men­tioned a nice big grant from Google, and these things are ongoing. 

Intertitle: Describe one of the break­through moments or move­ments of the Internet in which you have been a key participant.

Goldstein: Well a lot of the break­throughs you know, were tech­ni­cal, and I did­n’t have a part in that. For exam­ple, a huge break­through was the Border Gateway Protocol, which allowed net­works which would had a lot of com­pli­ca­tion inter­nal­ly to avoid hav­ing to exchange all this detail by just hav­ing con­nec­tions among these net­works at bor­der gate­ways. And Border Gateway Protocol I believe is still work­ing quite active­ly in the glob­al Internet. 

One of the prob­lems is that… Let’s see, the struc­ture of net­works is such that if you have a lot of very small net—rel­a­tive­ly small net­works and you want to reach clients on them, the num­ber net­works just explodes to the point that rout­ing becomes a very com­pli­cat­ed mat­ter. And it would be dif­fi­cult to have routers, at least in the big-boy com­mer­cial world, that were built on PCs as they used to be built back in the 1980s and 1990s. And so I would…as a lay­man on this, and I actu­al­ly am a lay­man when it comes to this, I say that some of the routers are approach­ing super­com­put­er size of the 1990s, not maybe today’s super­com­put­ers, in order to be able to work. They’re amaz­ing things. 

And now because we’ve pret­ty well run out of IP ver­sion 4 address­es, peo­ple are reselling blocks of address­es that they used to have and they’re chop­ping them up. Well, that makes rout­ing even more dif­fi­cult. So the ques­tion is to what extent will reach­a­bil­i­ty of indi­vid­ual IP address­es be main­tained? And I think that’s even a prob­lem today. It’s only going to get worse. 

Intertitle: Describe the state of the Internet today with a weath­er anal­o­gy and explain why.

Goldstein: That would be a very sim­plis­tic char­ac­ter­i­za­tion. The way I did talk about is this. There are days which are sort of part­ly cloudy, part­ly sun­ny, but they say well but, pos­si­ble thun­der­storms late this after­noon. Some of which could be dan­ger­ous, could even spawn tor­na­does. And that’s the way I’d sor­ta char­ac­ter­ize it. Because the Internet kind of bum­bles along now. It’s held togeth­er some­times with bal­ing wire and chew­ing gum. Because of all these rout­ing prob­lems. Because of links that go down and paths that have to be redis­cov­ered and so forth. Because of third- and fourth-tier net­works that need to get rout­ed in the net­work but have trou­ble get­ting a peer­ing rela­tion­ship with the big boys. And then all of a sud­den you get hack­ing inci­dents and trou­ble­mak­ing and so forth. So, the thun­der­storms come in, the light­ning comes in. 

And then we learned that gov­ern­ments are real­ly snoop­ing on all the stuff that’s going on and so that casts big, gray clouds over the whole thing. And to what extent will the Internet be able to per­sist as a glob­al medi­um. Will the weight of all these things be sup­port­ed by Internet tech­nol­o­gy. And I don’t know the answer for that one. 

Intertitle: What are your great­est hopes and fears for the future of the Internet?

Goldstein: Let’s put it this way. When the Internet got start­ed, it was large­ly an aca­d­e­m­ic and research under­tak­ing. And it was based on Unix. And there was real­ly no real secu­ri­ty built, oth­er than things like pass­words and maybe some encryp­tion here or there. And the atti­tude that my boss had at National Science Foundation was, That’s not our con­cern. This is for the aca­d­e­mics. People want to build in all kinds of secu­ri­ty, that’s some­body else’s problem.”

Well…it’s…I think that was a very valid point at that time, but that was 1990. And secu­ri­ty… How do you build secu­ri­ty on a sys­tem which was nev­er designed as a secure sys­tem? That’s the ques­tion. There are sup­pos­ed­ly secure gov­ern­ment net­works in var­i­ous coun­tries. How secure they are and how they work, I don’t know. I know they exist. Would those secu­ri­ty mod­els be gen­er­al­iz­able to the pub­lic? I don’t know. My guess is no. Will a whole new archi­tec­ture have to be built to han­dle secu­ri­ty? I’m not sure. But my guess is no mat­ter what you build, some­body’s going to be able to hack into it. And no gov­ern­ment, or no big gov­ern­ment, is going to let peo­ple build net­works that don’t have back­doors to them. We’ve learned in the head­lines that gov­ern­ment is ask­ing Internet access providers to build back­doors into what­ev­er they offer to the public. 

So that’s very dis­turb­ing. I under­stand it from the point of view of gov­ern­ment want­i­ng to keep track of the bad guys. But for the casu­al, the every­day user, it’s rather dis­turb­ing to know that his gov­ern­ment and oth­er gov­ern­ments can real­ly fig­ure out what he’s doing and then read his or her mail.


Yeah, the oth­er con­cern that I have, and I have no idea where it’s going, is the idea of con­sol­i­da­tion and com­mer­cial­iza­tion. So, in the ear­li­er days we had a lot of inde­pen­dent Internet ser­vice providers—ISPs. And grad­u­al­ly they were bought out as the indus­try con­sol­i­dat­ed. But the indus­try con­tin­ued to con­sol­i­date. And grad­u­al­ly com­pa­nies that at that point we said were total­ly clue­less… Like AT&T. AT&T was total­ly clue­less when it came to Internet in the 1990s. And every attempt we made to involve AT&T in the Internet, they kin­da turned their back and they weren’t inter­est­ed. So that’s how MCI and Sprint, for exam­ple, got inter­est­ed. But AT&T was­n’t ter­ri­bly interested.

Now, of course AT&T has under­gone cor­po­rate changes and the merge with Bell South— Was it Bell South or Southern Bell, I for­get which. So they’re dif­fer­ent cor­po­ra­tion now than they used to be. But here you have com­pa­nies that’re basi­cal­ly com­mu­ni­ca­tions com­pa­nies that are now run­ning large seg­ments of the Internet. And I’m sure their phi­los­o­phy is basi­cal­ly a phi­los­o­phy of rev­enue and rev­enue gen­er­a­tion for their share­hold­ers and have very lit­tle sen­ti­ment for the Internet the way we viewed the Internet when it first got start­ed. They’re just look­ing at it as a way to again, gen­er­ate rev­enue for their share­hold­ers by pro­vid­ing a ser­vice. Where they will go with that I don’t know, but I don’t know to what extent the inno­va­tion which con­tin­ues to evolve with­in the aca­d­e­m­ic and research com­mu­ni­ties will be trans­ferred to the com­mer­cial Internet providers. 


Well, there’s cer­tain­ly a lot of devel­op­ments again with­in the research com­mu­ni­ty. Especially things which are sup­port­ing what I call computationally-intensive research. Now, to what extent those broad­band things are going to be even use­ful to the aver­age con­sumer oth­er than for enter­tain­ment pur­pos­es like watch­ing stream­ing video, I’m not sure. So there is some excit­ing stuff, but I don’t know to what extent it’s going to apply. 

On the oth­er hand there are a lot of use­ful things that are going on, appli­ca­tions and gad­gets that go along with, or… Gadgets maybe not be the best word but…what should we call them? Devices that go along. And I can give you an exam­ple of one that I find par­tic­u­lar­ly use­ful but it’s just one. And he’s search­ing into his pock­et now for…ta da! This belongs to me but it’s not my devel­op­ment. These are actu­al­ly elec­trodes for an elec­tro­car­dio­graph. And this is just a phone, a case that fits on an iPhone. So it fits over an iPhone. And with this thing, I can take a lead one—I think it’s lead one or lead two elec­tro­car­dio­gram. And turn it into a PDF and send it to a physi­cian for inter­pre­ta­tion. By email. And that’s just one exam­ple of mobile med­ical devices that say hook onto a smart­phone. There are oth­ers for example…some peo­ple at MIT I think are on the verge of com­mer­cial­iz­ing it, some­thing that can be used in the third world where they hook Something to the iPhone and peo­ple do things to sort of match pat­terns, get pat­terns to con­verge. And from this, they can get a pre­scrip­tion for eye­glass­es. And the next step is that that pre­scrip­tion can be sent over the net­work to some­body who can grind a pair of glass­es and send them back to the user. Or to the per­son. This could be use­ful in many parts of the United States and not just the third world. 

So those are just some of the things, devices, which’re using Internet tech­nol­o­gy which I think are bright points. 

Interviewer: That is so inter­est­ing. I’ve nev­er [inaudi­ble; crosstalk]

Goldstein: Oh you should try it, if you want. 

Interviewer: [laughs] That’s really…oh, that’s so cool. I’ve heard a lit­tle bit about some of the [inaudi­ble; crosstalk]— 

Goldstein: Yeah well actu­al­ly I’m devel­op­ing a brief­ing to give to a group of senior cit­i­zens on mobile med­ical devices so. I mean you’ve heard of Fitbits.

Interviewer: Mm.

Goldstein: Okay, so Fitbits are—there are a lot of things like Fitibts. Nike has one. There’s just you know, per­son­al kinds of things. There’s actu­al­ly an appli­ca­tion on the iPhone but I’m sure Android has it too, where you can put your fin­ger over the cam­era and it’ll take your pulse, tell you what your pulse rate is. That’s not all that excit­ing. I think the EKG is even more exciting.

Interviewer: Yeah. That’s real­ly cool. There’s a few hos­pi­tals in the area where we’re from that’ve kind of start­ed a large-scale roll­out of some of those technologies.

Goldstein: Yeah. I mean, there’s one where you can get remote mon­i­tor­ing of a patient who’s say a bedrid­den patient by putting sen­sors on them and then it gets sent wire­less­ly, but it can be picked up on an iPhone by a physi­cian some­where else on the net. So you know, a lot of these things are devel­op­ing using the net­work. But to what extent—I mean are these rev­o­lu­tion­ary new tech­nolo­gies? I think they’re just things that learn how to use the net­work the way it is. I’m not sure where it’s going in the future.

Intertitle: Is there action that should be tak­en to ensure the best pos­si­ble future?

Goldstein: Yeah. I think that’s a polit­i­cal ques­tion. And right now, cer­tain­ly in the United States…political progress is pret­ty much at a stand­still. You know, Congress is dead­locked. I think the United States is not the only place where that is hap­pen­ing. So will peo­ple take action polit­i­cal­ly to curb or to lim­it some of these abus­es? I don’t know. I cer­tain­ly hope so. 

I can’t remem­ber who was the tar­get, but I imag­ine there were many tar­gets. Sometime I think…well in our past his­to­ry, sup­pos­ed­ly our spy agen­cies con­spired to get urine sam­ples from dic­ta­tors, or from poten­tial adver­saries so they could ana­lyze the urine sam­ples and infer what the per­son­’s health was and how long this per­son was like­ly to live. Okay. That’s real. Or at lest I’ve read it I believe it, okay. 

Well, let’s say that some­body was send­ing med­ical infor­ma­tion over the net­work. And let’s say that some­body that was send­ing med­ical infor­ma­tion­al, or some­body whose med­ical infor­ma­tion was being sent, was a polit­i­cal leader. And so you would­n’t have to sneak in and get urine sam­ples out of the bath­room. I mean you, could actu­al­ly get real stuff which is being sent. And to what extent can that be pro­tect­ed? To what extent can a deter­mined hack­er hack into med­ical records and get stuff on say the President or the Speaker of the House and so forth. The Speaker the House has been absent for the last three days. Nobody knows where he is. Hey, let’s hack into his med­ical records and find out what’s going on. 

And that was a hypo­thet­i­cal case, by the way, you get that. 

Intertitle: Is there any­thing else you would like to add?

Goldstein: Yeah, there’s one more thing that… It’s a recur­ring thing that comes up, you know sort of cycles every five or ten years. And I think as of late Vint Cerf has been bring­ing up this ques­tion again. And that is records, dig­i­tal data. Digital data that we have today…or even stuff on paper and stuff on film and that was the ques­tion. Stuff on tape, stuff on film and so forth. In ten years, how are we gonna be able to read it? In twen­ty years, how are we gonna be able to read it? If we get rid of all the paper now and dig­i­tize it, as libraries are being dig­i­tized and so forth, how do we store it? Will the for­mats be read­able in the future? Will the stor­age media per­sist? Will the infor­ma­tion dis­in­te­grate dig­i­tal­ly? Will be it backed up and recov­er­able? If there’s a dis­as­ter and—will we lose infor­ma­tion through dis­as­ters? Will we lose it because for­mats change? Will we lose it because media changes? 

Already appar­ent­ly we’re los­ing a lot of TV shows. I just saw some­thing on TV about that. Old TV shows which were taped, all the tapes are dis­in­te­grat­ing. And peo­ple are try­ing to pre­serve them and dig­i­tize them. So okay, so they’re dig­i­tized. But will the dig­i­tal libraries per­sist in terms of for­mat and stor­age and so forth twen­ty or thir­ty years from now? So that’s not so much an Internet ques­tion but it is an infor­ma­tion ques­tion. And of course the Internet is a medi­um by which infor­ma­tion is exchanged and stored, so that is a very impor­tant question.

Further Reference

Steve Goldstein pro­file, Internet Hall of Fame 2013


Help Support Open Transcripts

If you found this useful or interesting, please consider supporting the project monthly at Patreon or once via Cash App, or even just sharing the link. Thanks.