【OHI】Leonard Kleinrock interview selection
We were using a telephone network to create a new network which will displace the telephone network. The iron is beautiful. Okay, so we get ready, so we typed the L. Tony said, “You get the L?” Here said, “Yeah, I get the L.” Type O, “You get the O?” “Yep, I get the O.” Type G, “You get G?” Crash. The network is done, so the first observation is what was the first message on the internet. “Hello” as in “lo and behold”. Now we didn't plan that. You know other great communications activities found as the pioneers have really good messages. Alexander Graham Bell and telephone, “I need you”. Samuel Morse, with telegram “What hath God wrought?” Armstrong, “a giant leap for mankind.” They were smart. They understood public relations, they understand PR and the media. We want that's fine, but we came up with the shortest, most powerful, most prophetic message we could. “Hello” as in “lo and behold” by accident and that happened at October 29, 1969 at 10:30 at night.
We were using a telephone network to create a new network which will displace the telephone network. The iron is beautiful. Okay, so we get ready, so we typed the L. Tony said, “You get the L?” Here said, “Yeah, I get the L.” Type O, “You get the O?” “Yep, I get the O.” Type G, “You get G?” Crash. The network is done, so the first observation is what was the first message on the internet. “Hello” as in “lo and behold”. Now we didn't plan that. You know other great communications activities found as the pioneers have really good messages. Alexander Graham Bell and telephone, “I need you”. Samuel Morse, with telegram “What hath God wrought?” Armstrong, “a giant leap for mankind.” They were smart. They understood public relations, they understand PR and the media. We want that's fine, but we came up with the shortest, most powerful, most prophetic message we could. “Hello” as in “lo and behold” by accident and that happened at October 29, 1969 at 10:30 at night.
Leonard Kleinrock
The first message
We were using a telephone network to create a new network which will displace the telephone network. The iron is beautiful. Okay, so we get ready, so we typed the L. Tony said, “You get the L?” Here said, “Yeah, I get the L.” Type O, “You get the O?” “Yep, I get the O.” Type G, “You get G?” Crash. The network is done, so the first observation is what was the first message on the internet. “Hello” as in “lo and behold”. Now we didn't plan that. You know other great communications activities found as the pioneers have really good messages. Alexander Graham Bell and telephone, “I need you”. Samuel Morse, with telegram “What hath God wrought?” Armstrong, “a giant leap for mankind.” They were smart. They understood public relations, they understand PR and the media. We want that's fine, but we came up with the shortest, most powerful, most prophetic message we could. “Hello” as in “lo and behold” by accident and that happened at October 29, 1969 at 10:30 at night.
The origin of the engineer's dream
And I was reading then said this is interesting. I can build this thing. It’ll cost me nothing. And I’ll hear music tune in radio stations. Let's build it like fine. And what I needed was things I could find around the house in the street, things like an empty toilet paper roll, and which I had to wind some wire to make what's called an inductor.
The wire I could find in the street, I needed basically a crystal. It’s crystal radio. And you can make that out of your father's razor blade on a piece of pencil lead. And I needed an earphone. I didn't have one, but I know that the candy store down in the street had a telephone booth. And I knew you could unscrew the earphone. So I stole the earphone. And then it needed something called a variable capacitor. One of these things that would allow you to tune it. And I didn't have that.
I know I couldn't find that in the street, but I knew a way to get one. I knew that downtown Manhattan was a street called Canal street, another one called Cortland street. All the electronics that it was coming out of World War II, um, just before lot of.
So my mother took me on the subway down there. I walked up to the first store I could find, and I bang my fist on the table, said, “I need a variable capacitor.” And the man behind the counter said, “What size?” It blew my cover. I had no idea. So I told him what I was wanted and fully knew exactly what I needed. He sold it to me for a nickel five cents.
I came home and I put these things together. And I could certainly hear music. And this was fascinating. I had no idea how it worked. I said, I wanna find out what makes this work. And I’m still trying to find out. You know, electromagnetic is a mysterious world. We know the mathematics. We know the electronics electricity, but it's still a wonderful mystery.
So I very much enjoyed making the… (How old were you at that time?) I was probably six or seven years old. Um, it was uh, so challenging to do this. You know, I like to build little things. And once I built that, I didn't realize it, but I was hooked. I was now destined to become an electrical, but I wasn't.
Invention of queueing theory
And it was a big fight, as I understand, to get TCP problem into two pieces. A TCP does have this, well behavior where you get more throughput than you drop it in half. I don't know if you're familiar with the protocol. It’s very bursting in burpee. And the person who put in some control, congestion control functions was gonna invent Jacobsen. He put in what we now know is the TCP equation one.
Now, in 1979, I published a paper, 1978, and then another paper,1979. And I explained that the way to get efficient performance is not to let the queues build up, but to keep the minimal. The theory is this, here is queue and the server. Suppose you want to minimize the time it takes to get through here and still keep the server very busy. If you could control how you put people in, what would you do? You put one person in and let him get served when he leaves, put another one in. Don't put people in the queue because that's slowing things down. And the engineering concept there is keep the pipe just full. If you have many cups, keep one in each cup. Don't flood it with a lot of other stuff that gets in the way in buffers and what's called buffer block.
I published that. Nothing happened to it. Last year, Brian Jacobsen looked at the paper again, said, “that's the way to do it”. And he's working for Google now. And one of their data centers just published a paper saying this, just keep the pipe wasn’t the way to do it. And they're getting rid of this kind of behavior now. So, I mentioned TCP, this is going to replace TCP, I believe.
