Oct 30, 2011

Interview with Mr. Vinton Grey Cerf - a.k.a. “The Father of the Net”

Mr. Cerf, a man who I have read about in several books during my computer network studies, the man whose findings and researches ended up creating nothing less than the Internet, well, seemed a lot improbable. But with a little persistency and a LOT of good will of Mr. Cerf, this interview was materialized. Thank you Mr. Cerf, for being such an accessible and kind man. It was a great honor!

To all: Enjoy it!

Best regards from Brazil,

Marco A. Filippetti

Mr. Cerf, first of all, on behalf of all readers of this blog, I would like to thank you very much for making time on your amazingly busy schedule to participate in this short interview. As I said before, it is an indescribable honor to have the opportunity to interview a living legend, the man who fathered the Internet.

<——- Interview Starts ——->

FILIPPETTI: At late 60’s you participated at the startup of the ARPANet, interconnecting its first two nodes. Many readers of this blog - me included - would like to know:

FILIPPETTI: Did you imagine, at the time, that the ARPANet and the TCP/IP protocols would end up affecting billions of people as it evolved to the Internet?

MR. CERF: At the time I was working on the host protocols for the ARPANET as a member of the Network Working Group team led by Steve Crocker, I think I had no idea what might happen. I was at UCLA at that time (1967-1972) and also working on a Ph.D. as well as serving as principal programmer for Prof. Leonard Kleinrock’s Network Measurement Center. In 1973, Bob Kahn and I began to work on the design of an architecture and protocol suite that would allow arbitrarily large numbers of packet networks to be interconnected. At the time we were doing this work, we were well aware of the applications that were possible, including things like packet voice and video. We were also well aware of the work of Douglas Engelbart at SRI International to support collaborative knowledge work with mouse, portrait-mode black/white displays, hyperlinking and powerful editing and document sharing tools. While we probably did not foresee at the time the global adoption of this technology, we did know that it provided a very potent and extraordinary opportunity to allow a broad range of invention by a wide range of contributors. It was an extremely open, distributed design and adopted a basic posture of accommodation for almost anyone with an idea to try it out. Our idea was that if you could implement the protocols you could connect to the network and become a part of it.

FILIPPETTI: What were the most challenging issues faced at that time?

quote2.gifMR. CERF: When the design work was first under way, we needed to invent a way to refer to different networks and to give addresses to each computer on each network. That led to the Internet Protocol (ultimately version 4). We also needed to develop the concept of a “gateway” that could relay Internet packets from one network to another. We did NOT translate packet formats between connected networks. Rather, we encapsulated each Internet packet in the packet format of each network. This is sort of like putting a postcard into an envelope specific to each connected network. Eventually, gateways became routers and everyone spoke “IP” even in the core of the network. Since then other networking technologies like Multi-protocol Label Switching (MPLS), Frame Relay, Asynchronous Transfer Mode, varying kinds of wired and wireless Ethernets all encapsulate Internet packets for transport within each network.quote4.gifThe next big challenge was getting the protocol implemented on many operating systems of the time (about 30 I think). Then we needed to get everyone to actually implement the net protocols replacing the ARPANET Network Control Program (NCP). This was one 1 January 1983 (with a few laggards). After that, the battle was global standardization and we had to wrestle with X.25/X.75 and Opens Systems Interconnection standards that were rivals. Eventually, TCP/IP became the de facto standard. Then came the WWW around 1989 and the commercialization of the Internet itself. Many policy issues now face the Internet community as it becomes a global resource and infrastructure.

FILIPPETTI: Nowadays, it is common knowledge that IP address exhaustion is going on rapidly. Forecasts point that by 2013 there should be no more public routable IP address blocks available. The version 6 of the IP protocol came as a response to that, little more than 10 years ago (in 1998). Still, it can be said that IPv6 is only in its infancy in terms of general worldwide deployment, having little more than one percent of Internet-enabled hosts in any country. That leads to the question:

FILIPPETTI: How do you see this issue (of IPv6 worldwide deployment) summed up with the rapidly depletion of IPv4 usable IP addresses?

quote5.gifMR. CERF: Generally, I think pressure will build as the runout becomes most visible in mid 2011. I see evidence of implementation of IPv6 in key places (and proud to say that Google has done its homework in this area). There will be hacks to deal with IPv4 and IPv6 inability to interact directly but there really isn’t any sensible alternative to implementing the larger address space.

FILIPPETTI: Now, with so little time to adapt, do you see a solution to minimize the impact to the end user and to the ISPs (regarding the v4 to v6 migration problem)?

MR. CERF: There are some steps using Network Address Translation tricks (see ComCast) that might help. However, most laptop and desktop operating systems, most server operating systems and most routers have the IPv6 code – they have to turn it on and learn to operate dual-mode networks, servers and clients. It is not trivial but also not hard – but it does take patience to get the details right.

FILIPPETTI: Technology is a living thing. It rapidly changes the world as we know by the creation of new ways to interact with it.

FILIPPETTI: What do you see coming as the five or six concepts or technologies that might change the world, as the TCP/IP did?

MR. CERF: Wireless for sure (and ultra-broadband); Delay and Disruption Tolerant Networking (DTN) that is being proposed for interplanetary communication and noisy/hostile mobile operation; non-fossil fuel sources (various kinds of solar; geo-thermal); LEDs to replace conventional light sources; patient-specific medical treatments (driven by genome and proteome information); neural-electronics (such as cochlar, ocular and spinal implants); quantum computation for certain kinds of computations.

FILIPPETTI: To wrap up this brief interview, Mr. Cerf, our readers would like to know: You are working at Google now, one of the most important and visionary technology companies in the world.

FILIPPETTI: How did you end up working there?

MR. CERF: I sent an email to Eric Schmidt [Chairman of the Board and Executive Director @ Google] asking if he needed any help and he said “yes”.

FILIPPETTI: What is your role in this great company? Can you mention any revolutionary – non-confidential - project you are working right now?

quote6.gifMR. CERF: We don’t discuss projects we are not ready to announce so I can’t tell you anything secret. We are working hard to improve ways in which people can interact with network resources. Our recent announcement of automatic captioning of English language video is an example. We are very focused on improving accessibility of our products and services. I am very interested in the confluence of all media onto the Internet and particularly curious to see what implications this has for entertainment and education. I am also interested in new modes of interaction (speech, gesture) in addition to keyboard and mouse paradigms.We continue to look for more ways to help organize information including our own medical information in a personal health database.

<——- Interview Ends ——->

Mr. Cerf, once again, thank you for accepting this invitation! I still can’t believe I am exchanging an e-mail through the Internet with one of the persons who created it :-D

My best regards,

Marco A. Filippetti


No comments:

Post a Comment