Episode 9: Interview with Abhay Bhushan, author of File Transfer Protocol

September 15, 2017


Pramod Shashidhara

Abhay Bhushan is one of the early pioneers of the internet. He’s the inventor of the File Transfer Protocol (FTP);  an application layer protocol of the Internet protocol suite, and is responsible for transferring computer files between Internet hosts. He has been a major contributor to the development of the Internet TCP/IP architecture, the early versions of email protocols. He is currently chairman of Asquare Inc. and President of the IIT-Kanpur Foundation.


Pramod: Hello everyone, welcome to mapping the journey. When we think of Internet and Internet protocols, it’s HTTP, TCP, FTP, SMTP to name a few. We have been studying these protocols for many years. Among these, FTP made the transfer of files from one computer to another computer possible. I always wondered, how these early engineers thought through and authored these protocols. Today, I have an opportunity to speak to the author of file transfer protocol, Mr. Abhay Bhushan. He’s is a pioneer in the field of Internet technology. He’s has been a major contributor to the development of Internet TCP/IP architecture and the early versions of protocols. He is mentoring startups today and also chairman of Asquare Inc.

Pramod: Abhay Bhushan, with immense pleasure and honor, I welcome you to the show.

Abhay: Thank you Pramod.

Pramod: Please introduce yourself to the listeners.

Abhay: I Abhay Bhushan, grew up in India and was in the very first batch of IIT Kanpur, and started 1960, and we were very fortunate because we were funded by consortium with 8 US Universities, the top ones, MIT, Princeton, Caltech, Berkeley, Carnegie Mellon, Universities of Michigan, Ohio State, Purdue. So there were these eight universities, from which professors visited IIT. So as a result, we had the very first computer center in any educational institute in India.

And the rumor was, it was brought in bullock-carts. I saw it personally and I said, it was on a flatbed with wheels. It could have been put by Bulls but people in the last journey, it was brought by truck. I didn’t see a bull pulling it but it said it was brought up on a bullock cart, so that’s a little exaggeration, so it’s kind of interesting because in 1962, and Harold Huskey just recently passed away was considered the kind of real pioneer in computer science, he is the one who came and brought the different computers, and we had some of the most brilliant people, for example, 1965, my mentor, who was there from 1964 to 1966, William Schreiber from MIT, he brought the first, very first television set there, which means people could produce TV programs and broadcast them. So we’re broadcasting it and the Government of India said, it’s illegal broadcasting so he put cables under the hostels in different places, so we had cable TV at IIT Kanpur 1965.

Pramod: 1965, that’s amazing.

Abhay: Yeah that’s private, our first convocation in 1965 was broadcast on TV and people could see it from different places, not just the Convocation Hall, where it took place. You know you can imagine the kind of things, the wind tunnel, and the only airfield in any Institute in India was at IIT Kanpur.

Pramod: Life after graduation from IIT Kanpur

So in 1965 after graduating, I came to MIT, and there was no Computer Science Department. The Computer Science Department all over the world including MIT started in the 1980s. All part of Electrical Engineering because that’s where the computers came from, so that was the original department so, but we did computer science and other things.

And, after doing my Masters and in fact for my masters’ thesis, I went to Bell Labs and worked on the transmission of color pictures and picture phones. So you know you see all those now handheld devices where you can video talk to somebody else, it was there in picture phones in 1965 during the world fair in New York City they were demonstrating it. So I saw that I said okay, I’ll apply to the bell lab people. They said yeah, this is all black and white, 6.3 megabit, big lines you know, it was not practical because the bandwidth was not there. AT&T Bell Labs I mean, and then I’m gonna come to the internet development related to this. So, I was doing like efficient transmission encoding, really encoding so how can you get the bits? The smallest number that we’ll be able to transmit, the color signals so that it will work on a 6.3 megabit picture phone link and so encoding becomes an important part of it.

Pramod: Life after graduation from MIT

Abhay: And so after graduation, I said I want to work and so I took a job with the electronic system lab, part of project Mac, project Mac is where the Multics, the multi-access computing, the first time-sharing system CTSS was done at MIT and so there I was thinking you know, if we’re doing something and we have all these minicomputers, can’t we link them together and develop some transmission protocols, so we can interconnect so I wrote a paper in the 1968 spring joint computer conference,  Inter Computer Communication.vSo, one of the professors who were there, professor Licklider, who was considered the father of the ARPANET. He’s reviewed a book called “libraries of the future”, that all the information will be online and people will be able to access it. The storage was very expensive of course in the 1960s, So he called me and he said, “you know very interesting, I’m also thinking of getting this thing funded by ARPA”, he said why don’t you come with me to the Pentagon, end of 1968 or early 1969.

So I traveled with him to the Pentagon and they were trying to write the specs for the bids going out for developing a packet communication network on which the different research computers from all ARPA sponsored research projects, it was ARPA then and DARPA now. So, I went there with him and you know they were focusing on security, so I asked the question, is it because you don’t want the Russians to know what we’re doing? No he said we want, the guy said, we do know, he was from the air forces, we don’t know an army to know what the Air Force is doing and the Air Force know what the army is doing, ok it’s kind of very interesting, it was a funny comment, I thought that’s very funny.

So we built the specs, I wasn’t quite involved, I reviewed them, Licklider reviewed them. Beranek and Newman won the contract, and so they developed this word called “IMP”, Interface Message Processor, so the IMPS were connecting with package switching Network all over the different places, so every kind of place where the computers were connected, there was an IMP (interface message processor), so that’s the one that was interfacing the packet communication network which they had developed to the computer’s.

Pramod: That is like our today’s Internet?

Abhay: Yeah it’s, it’s like today’s, well it’s precursor. The Internet comes later, I’ll come to why it’s called the Internet, so that’s called ARPANET, at that time. So 1969, is when they’ve, the contract was launched and they developed these IMPS and 1969 is the first time they connected three nodes together, now the speeds, the communication speed between these IMPS that were connecting it were, maximum was 256 kilobytes per second, now we think of megabits and gigabits and gigabytes, and 256 was considered a very high-speed line and some other ones like 56 kilobits per second, that’s how at the rate at which data was found.

My friend Bob Metcalfe, who you might know, who’s the developer of the Ethernet, he’s best known for that and 3com corporation, so he and I were kind of in this office next door so he said I’m gonna connect the PDP-10, so my choice was to connect the Multics system, the big G-645 Multics System so I developed the interface. For this, we had a little bit of a friendly race as to who gets to finish it first.  So the, my hardware is finished first but the software was taking longer and his software PDP-10 was less complex and so he got the whole thing working first, I got the hardware working first, so he said that’s a tie, okay,  so Bob is a good friend from MIT days and from Xerox days.

And, so at the same time, I also applied for Ph.D. program, was in the PhD program in 1968 and I wrote my thesis as a “resource sharing networks” and my thesis committee said, “well that’s an engineering project, it doesn’t quite qualify for research, we don’t see what the research element for Ph.D. is”. So I said, “well I’m gonna work on this okay”. So I dropped out and I got an MBA from MIT Sloan school of management. My whole idea was to go back and do something, build something as opposed to being academic and teaching and so that’s just my own personal bent.

Pramod: Invention of File Transfer Protocol

So we formed a small working Network working group in 69, late 69 because I said now what do we do, how do we communicate the computers to each other, we have this packet flowing through, we can send packets, but how do you make sense of it, that packet flowing from here, computer A to computer B, like UCLA was one of them, Vint Cerf was at UCLA graduate student and Steve Crocker, so that was the first computer then Newman is the third computer and I, I think MIT was number four or number five to get on the network, so we all, like five places and more sites were coming in gradually so he formed a first Network working group, he was the chairman of the network working group.

Then you know, we were developing TCP/IP, packets and all those kind of things that you use commonly even today and say okay, how do you find out whether this packet and how do you assemble packets going from different places and how do you know which belongs to what, so that’s the origin of the TCP/IP so, it was actually called transmission control protocol, Internet Protocol came later which I’m going to come back to, it comes about five-six years later, , okay or four years later and so at that point we said okay, we first caught the thing, computer’s communicating, okay so we can send a message from “A” computer to message to “B” computer and knows what the message is getting intact and assemble. , that’s the first part of it, the second part is, kind of thing okay, now what do we do with it? , okay, at that time we had would you believe acoustic couplers, have you heard of them? (no), your telephone, everybody had the telephone.  so there were these devices, you put the telephone on and then you could use the “teletype” like device and send messages and connect your computer,  from your home, little small, luggable teletypes are available either on my home and then you take the phone and you put that acoustic coupler, put the acoustic coupler to the teletype device and now I am, I can login to my computer.

Pramod: Just like a shell access of today?

Abhay: Yeah I mean so those things were there so now you say okay, first thing is, I can log into my own computer, you could talk to my terminal, so these are terminals connected from the home telephone networks from an acoustic coupler. So I want to use the UCLA computer; they give me a password, so first protocol developed was Telnet protocol.

That means now you can communicate, use one terminal connected to say Multics, use the PDP-10 from it directly, without connecting to PDP-10, so that was Telnet. So we developed that, I was part of the Telnet team,

And then another one, I said, “no, we really need more applications ok.” What do we need? We use email and we use files and transferring the files within a system,  so why don’t we get work on both these applications? so they said, “well, that’s a good idea, why don’t you become a chairman of the application, File Transfer Protocol?”

So I said ok, so we got you know, it’s all developed in a team, one person is just kind of you know, when I say I wrote the thing, I wrote the protocol, that was really a team effort, as it always is. We build on ideas of others, we do the things, somebody gets to document it, somebody gets to lead it, drive it, see what’s being done, other people up implementing, writing the software, they find some bugs, they said no no, this doesn’t work, let’s do it this way.

So we developed FTP, so I wrote the first protocol in April of 71, it takes about 6-8 months but yeah, at times you’ve got this going, that going, so the first time it came out and then you know we can, keep different versions, kept improving it, The last one I was in like 356 or 561 is the last one I wrote the version and then I left in to join “Xerox”

But the interesting thing is, at that point we said we want to do email  and if you really recall those days, one megabit or two megabit was a mainframe, big computers were  two megabit core memory , many of them were like 256 kilobyte core memory, now you know you have this gigabyte, these core memories over here, you can’t imagine how, why you were not to make things efficient, so we said we don’t want 2 processors, that’ll be too taxing on some other like one megabit machines or 500 kilobit machines so we said why don’t you put two commands into FTP called mail and mail file,  so email is like normal text messages, mail file is mailing attachments, what you have today,

Now, How do we identify? so we have a name, you have a username and at “hostname”,  so we had username a “space” AT “space” hostname, so we need to standardize network headers. Ray Tomlinson, who just recently passed away, he was at Bolt Beranek and Newman and he was using 3-Bolt Beranek and Newman computers “@” sign to differentiate between the username and the hostname, that, they’ve three localhost, so I said you know Ray, that’s a nice thing, you know rather than “space” and “AT” and people have to write programs to parse space and an “AT” and a space and then the name, why don’t we just use a single symbol, he said that’s a great idea, so we were in a group, Ray was part of it, Bob was part of it, myself, so we wrote the “RFC”, you can look at it, “Standardizing Network Headers”, and then we put that, we’re all going to use “@” sign and we want to use what is allowed in the network names and “@” sign is not used in anybody’s name. Sometimes people are using periods (.), but they’re using other things but nobody is using “@” sign, they’re using dashes (-), you know, some people allowed that some computers, some people but no spaces were also not allowed in the names at that time and still they’re not. You cannot use space. (yes, we cannot), right, so we said okay, that sounds good hmm, that’s a good punctuation to distinguish the two things,

The only problem was the teletypes, “@” sign was a kill character,  a null character, no control outside so he said, but that teletypes are not that important, you know you can always if you’re connected, a computer connected to teletype, they can always substitute,  so it worked, (awesome), and that was the way the network standards got headed, yeah and then after, formed an SMTP and then IMAP and all came much later.

Abhay: Yeah so, it’s kind of you know the things developed, the FTP has improved tremendously now we have Hadoop and the high-performance things and you know big data and we can transfer very high speeds and your Google with the world’s information and you’re transporting very high speed, your video transmission, better encoding and better kind of a things, so this encoding makes it possible for us to transfer it even to cell phones, video images are pretty high quality, movies to televisions and so you know, all this is kind of development shared by a lot of people, and you know we all walk on the shoulder of giants who came before like Hareld Huskey, like people who were real pioneers, whenever Bush at MIT, they were the real pioneers who developed the computers and other things and building blocks and, so you build up upon that,  and after those hundreds of people, did brilliant work to build upon it.

Pramod: Yes, the building block was laid by you all, and I never knew, like email and FTP, they were, they were together, I did not know that.

Abhay: Yeah for about four years or so.

Pramod: Okay. So you could use mail to send and receive …

Abhay: Yeah yeah for the FTP because they only wanted one server, another thing I did was a survey program which is the precursor, some of the bots that collect information and things so we said okay, we want to know how much each computer is up, how much data is going through, so I wrote a little survey program which will ping different computers and see they’re up or not and then maintain statistics kind of fun little thing to do.

Pramod: Okay Abhay I have to interrupt you know, you told us the history of all the major protocols wonderfully well and you are involved in the development, we’re so glad to be speaking with you today. Thank you and please continue.

Abhay: And Then when I came to Xerox okay, Xerox was the one who was developing all these models, my job where I’ve been to Rochester first because they said you know, there was a, do much more research and things so he said, “yeah well you know we need you”. They interviewed me here as well as in Rochester, they said, one, why don’t you bring these people up to the modern age you know because Xerox is developing you want somebody for the information technology to drive that,  and people just couldn’t understand it. It’s gonna tough so then, after a few years I took it one year off and to India to work in the villages as a personal passion  from my grandparents over freedom fighters and working for those so they said, they, when I came, they said you must spend one year working for the poor.

So I said this is my time to go. So in 78, 79 and I came back and I started working when the strategy for Xerox to bring into the modern computing world like with the personal computers and workstations and things and so, then I joined the system development division  where we were kind of you know, putting all the systems so, in 1980, we introduced the PC for Xerox and in 81, we used a star computer with completing network expected today’s internet and then Steve Jobs as you know, came to Xerox and got the Macintosh  things started, but we had the big star workstation all networked with the file servers, mail servers, print servers everything interconnected and a lot of people using it, white house was using it, Microsoft was using it and then Xerox flubbed it.

And there’s a book written called fumbling the future and what the Xerox did is the management did not understand. I was in meetings there and they said they couldn’t understand a bit from a byte, and they said hey how does, why does it take you to Sales, the salespeople. Six months to sell a system or a year. Our copy duplicators sells a big copy like a thing in two weeks, they couldn’t understand (understand the value), and the, the salespeople got discouraged, many of them left and joined Apple, Wang, IBM, many other companies and things, they just couldn’t  understand it and they were charging $10,000 or $5,000 for just administrative fee for the workstation  so you know it could have made it much cheaper and taken over the market but they just didn’t do it

And they finally sold their workstation division to Sun  in 87-88 so at that point I was saying, you know what should we do, we had, we still had laser printers in 97 hundred or things so at that point I was bicycling to work and I said you know, got involved in Earth Day 1990, settled a proposal to Xerox to green Xerox Corporation, to the CEO, he said, “great, you saved me money, what do you need?”. So I said ok give me my salary and then  a little budget and must have saved the company billions of dollars  just by reducing waste, it’s just like following the passion, so if somebody asks me what do we do to do something, follow your passion   for this technology, where there’s social improvement, development or environment, follow your passion. Be happy… It’s not about making money, (not at all). It’s about making a difference.

Pramod: This is one question which I really wanted to ask when you were developing those protocols, you had a vision of internet, so what do you think the current state like how, how has it changed over the years?

Abhay: Yeah, okay. We had a pretty good vision of the information being available, but none of us, because you and the academic area are not common, thought of commercialization of the internet, so what was really surprising in the 90s when I saw, was the huge way, the internet were commercialized , for sales of all goods, now even groceries,  Amazon is getting into it and that’s something we had not even thought about, talked about in the 60s and the 70s , they’ll be used for commercial marketplace, now that’s the biggest use of the Internet today,  in terms of most lucrative use, we thought about you know information, connecting people, employment, this thing that how to make it easier, so only thinking information, not goods and services but if you really translate goods and services are really information, when you buy and sell its information. Then of course, you have, somebody UPS and someone, delivery will be drone later on or something will deliver it, real goods.

Pramod: Very well said sir, and how about your association with India Today?

Abhay: But I’ve been connected to India  and helping Grassroot because of the work I did in India, so a lot of grassroots projects is a, organizing Indians for collective action and we help, we’ve helped hundreds of different projects in grassroots level education, health care, innovation, sanitation, you know it has been made, environmental protection, different fields, all over India, in almost every state.  so that’s something that you need to stay connected. Then in 2002 and 2003, bunch of us got together and said went to IITs, when we started pan IIT organization to promote  the IIT brand at the same time helped IITians in different fields, , so Pan IIT was created and I was the first founding president for the Pan IIT, and that is me satisfying it to help our alma mater  like IIT Kanpur Foundation to help IIT Kanpur, the IIT Bombay Heritage Fund, so all these things got started you know at the same time, alumni here helping and now the bits and any other Institute they have their own now alumni network trying to help  so whichever place you get to so that you know these catch on and it’s good to give back, so giving back is very important.

Pramod: Awesome. Thank you Abhay, it was wonderful talking to you. It was my pleasure having you on the show.

Abhay: Thanks Pramod

Pramod: It was wonderful speaking to Mr. Abhay Bhushan, great personality it almost felt like I’m listening to a very interesting story on Internet Protocols. Contributions from such brilliant engineers made the field of computer science, they provided us the building blocks for engineers to build on today. Thank You, Abhay Bhushan.

And Thank you, listeners. Next episode I will be speaking Mr. Rob Das. Chief Architect and Founder of Splunk. Should be a very interesting conversation, I love Splunk and let’s ask him how he built this multinational company. Looking forward to speaking with him. Until Next time Adios

Leave a Reply

1 Comment on "Episode 9: Interview with Abhay Bhushan, author of File Transfer Protocol"

Notify of
Sort by:   newest | oldest | most voted
Ajit Dongre

Amazing interview! He is clearly one of the giants on whose shoulders others have walked to bring us the Web.