Radia Perlman: Thank you very much for this hon­or. I would like to say that I was born know­ing I was going to be design­ing pro­to­cols. But in fact, if I’d known more about com­put­ers when I was young, I would have said I would be hap­py doing pret­ty much any­thing as long as it didn’t involve com­put­ers. And I think that the fact that I real­ly am not in love with com­put­ers means that I design things for peo­ple like myself. I want things to just work and you shouldn’t have to under­stand it. Too much of the time, engi­neers design for oth­er engi­neers, and they ask you ques­tions like, Do you want to dis­play both the secure and the inse­cure items?” Which, I hap­pen to know what it means and it’s non­sen­si­cal. And you know, things just shouldn’t be that way.

So I very much believe that net­works should be com­plete­ly self‐organizing. And then, I was actu­al­ly pres­sured— You know, peo­ple told me, Well, some of our cus­tomers real­ly enjoy con­fig­ur­ing things.” Well, okay fine. You want to con­fig­ure things? So I put in knobs that you could play with, but you don’t have to touch the knobs. It’ll work with­out your touch­ing the knobs. And if you want to play with the knobs, fine. Any set­ting of the knobs will still work. So I believe things should be like, incred­i­bly easy to use because it’s actu­al­ly peo­ple run­ning it.

The oth­er thing is that we put up with just…junk. We come to believe that it’s of course nat­ur­al that your com­put­er will freeze every few days and you have to reboot it. That might be fine for a com­put­er. It’s not okay for the Internet. There’s no on/off switch for the Internet. So the Internet has to be absolute­ly robust. And so that was one of the oth­er things that I did, before I— I mean, one of the pieces of work that I did was to make the rout­ing of net­works what’s called self‐stabilizing,” which before that it was pos­si­ble that even though all of the routers were doing the right algo­rithm, if some his­to­ry of kind of bad pack­ets or one sick router would be emit­ted, the whole net­work would be down for­ev­er, even though the sick router that inject­ed it would be removed. So that’s real­ly impor­tant for it not to work that way.

And inter­est­ing­ly, in the paper in which I explained how the exist­ing stuff could get into this bad state and how to design it so that it would be self‐stabilizing, the last line I wrote in that paper was, Well you know, this is how to make it work. Once you get rid of the bad guy, the net­work will return to nor­mal oper­a­tion.” Well, when I went back to grad­u­ate school ten years lat­er, my man­ag­er at Digital, Tony Lauck, sug­gest­ed that for my the­sis I either prove my state­ment in this paper that you couldn’t pos­si­bly expect a net­work to work while the bad guy was still there, or fig­ure out how to do it.

And so my the­sis was actu­al­ly even extend­ing robust­ness to hav­ing a dis­trib­uted algo­rithm, or a net­work, work even when some of the par­tic­i­pants are mali­cious. So while they’re still con­nect­ed. And the big­ger the Internet is, we have to real­ize that there will some­times be mali­cious par­tic­i­pants. So when­ev­er there’s infor­ma­tion, a lot of it is going to be wrong and mali­cious­ly placed, and how can soci­ety still work? How can the net­work still work?

So the Internet has so many amaz­ing chal­lenges ahead. The basic tech­nol­o­gy of sort of get­ting it work­ing is almost dwarfed by the chal­lenges there are ahead of us. So it’s won­der­ful that any­body can say any­thing and dis­sem­i­nate it to the whole world. But then, how do you dis­tin­guish cor­rect infor­ma­tion from incor­rect infor­ma­tion? If there’s nobody sort of being able to tell you what’s true and what’s not true, it’s almost like why both­er look­ing it up? Just make up some­thing. The right stuff will be there, but a hun­dred times as much wrong stuff will be there. And so that’s a real dan­ger.

The oth­er thing is again, peo­ple don’t like cen­sor­ship, but what about gen­uine­ly dan­ger­ous things telling you how to cre­ate hor­ri­ble poi­sons or bombs and so forth? Do you real­ly want there to be no cen­sor­ship and to have infor­ma­tion like that be read­i­ly avail­able?

The oth­er thing is that infor­ma­tion should be free. It’s sort of so won­der­ful that some very poor vil­lager some­where can access these things, teach them­selves what­ev­er. But on the oth­er hand, if infor­ma­tion is free, so many news­pa­pers and mag­a­zines are now out of busi­ness. It costs a lot of mon­ey to actu­al­ly do inves­tiga­tive jour­nal­ism. And so if we expect this stuff to be free, where will the infor­ma­tion come from? Because some­body has to do that.

The chal­lenge of hav­ing all of these diverse cul­tures coop­er­at­ing togeth­er in the same Internet in terms of what should the rules be, what should be allowed to be dis­played, is…you know, amaz­ing­ly dif­fi­cult chal­lenges. But it is aston­ish­ing that the Internet works as well as it does today and has scaled as well as it does. But it’s cer­tain­ly a fas­ci­nat­ing time to be part of it. So thank you.


Help Support Open Transcripts

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