ARPAnet: History of the beginnings of the Internet, DARPA, RFC and other Network protocols

Every March, July and November, we are reminded that The Internet is not the mature and stable technology that it appears to be. We depend on the Internet as an essential tool for our economic, social, educational and political lives. But when he Internet Engineering Task Force meets every four months in an open conference that jumps from continent to continent, more than 1,000 people from all over the world meet with change in mind.

His vision of the global network shared by all humanity is dynamic, continually evolving and improving. His efforts are combined with the contributions of many others to ensure that Internet always worksbut never end, never complete.

The quick but neat internet evolution it is all the more remarkable considering how unusual it is produced: without a company, government, or board of directors in charge. Nothing about digital communications technology suggests that it must organize itself or, at any rate, be fundamentally reliable.

What is the origin of the Internet?

We enjoy an Internet that is both because several generations of network developers have adopted a principle and process that have been quite rare in the history of technology. The principle is that the protocols that govern the communication of devices connected to the Internet must be open, expandable and robust. And the process that invents and refines those protocols requires collaboration and a broad degree of consensus among all who care to participate.

Those individuals who were part of the small team that deliberately adopted a collaborative and consensus-based process to develop ARPANET protocolspredecessor of the Internet, are pleasantly surprised to see the persistence and success of those ideas, even as the physical network has evolved from the 50 kilobit per second phone lines from the mid-1960s to the connections of optical fiber, 5Gsatellite connections and submarine internet cables that we enjoy today.

Of course, technology advances. None of the computing or communication equipment that was used to build the ARPANET is today a crucial part of the Internet. But there is one technological system that has remained in constant use since 1969: the humble RFCwhich was invented to manage change itself in those early days.

DARPA and how an idea begins to take shape

The ARPANET was much simpler than the Internet because it was a single network., not a network of networks. But in 1966, when the Defense Advanced Research Projects Agency (DARPA) of the Pentagon When he started planning the idea of ​​connecting entirely different types of computers together at a dozen or more research universities from California to Massachusetts, the project seemed pretty ambitious.

It took two years to create the basic designwhich consisted of an initial subnet that would exchange data packets through dedicated telephone lines that would connect computers in only four places:

• The Santa Barbara campus
• The Los Angeles campus of the University of California
• The Stanford Research Institute (SRI) in Menlo Park, California
• The University of Utah in Salt Lake City

In each place, a router (we call them IMPs, for Interface Message Processors) chopped the outgoing bit blocks into smaller packets. IMPs also recomposed incoming packets from distant computers into blocks that the local “host” computer could process.

In the tumultuous summer of 1968, Steve Crocker I was a graduate student spending a few months in the computer science department at UCLA, studying there, Vint Cerf. Like many others in the field, Steve was much more interested in artificial intelligence and computer graphics than on networks. In fact, some principal investigators outside the top four sites initially viewed the ARPANET project as an intrusion rather than an opportunity.

When ARPA invited each of the four pilot sites to send two people to a kickoff meeting in Santa Barbara in late August, Vint and Steve drove from UCLA to discover that everyone else in attendance was also a graduate student or member. of the staff. No teacher had come.

Hardly any of them had ever met, let alone worked with, anyone from the other places before. But they had all worked on timesharing systems, which doled out chunks of processing time on mainframes to a series of remotely connected users. so everyone had a feeling that interesting things could be done by connecting distant computers and making your applications interact with each other.

In fact, they hoped that general-purpose computer networking would be so useful that it would eventually spread to include virtually all computers. But they certainly did not foresee how that meeting would set in motion a collaborative process that would turn this small network into a critical piece of global infrastructure.

After meeting in Santa Barbara, they set up follow-up meetings at each of the other sites so everyone had a common vision of what this eclectic network would look like. The SDS Sigma 7 computer at UCLA would connect to a DEC PDP-10 in Utah, an IBM System/360 in Santa Barbara, and an SDS 940 at SRI.

They would be a distributed team, writing software that would have to run on a diverse collection of machines and operating systems, some of which didn’t even use the same number of bits to represent characters. Adopting the name of the ARPA-appointed faculty committee that had assigned them this project, they called themselves Network Working Group.

Laying the foundations of the Internet

During the fall of 1968 and the winter of 1969, only a few months were available to complete the theoretical work on the general architecture of the protocols, while waiting for the R&D company Bolt, Beranek and Newman (BBN) build the IMPs in Cambridge, Massachusetts.

Through regular meetings, a broader vision emerged, made up of three ideas:

Firstseeing the potential of many network services with different application programs that could exchange messages over the network, and even remotely control and execute each other’s subroutines.

In second place, network services needed to be scalable. Timesharing systems had shown that a new service could be offered simply by writing a program and letting others use it. The network should have a similar capacity.

By latest, the network would be more useful if it were hardware agnostic of its hosts. Any software should be compatible with any machine without any problems, regardless of its word length, character set, instruction set, or architecture.

When the Network Working Group met in Utah in March 1969, drafting tasks were distributed to write down these 3 ideas mentioned above. Until the network worked, a email protocol, to share team memos via US mail. To make the process as easy and efficient as possible, a numbered list of documents in circulation was kept, and authors mailed copies of the Emails they wrote to everyone else.

The group of graduate students made their way through the darkness. They didn’t even have a leader. Surely at some point the “real experts” (probably from some renowned institution in the Northeast or Washington DC) would take over.

The RFC protocol, much more important than it seems

The first batch of RFCs came in April 1969. What was possibly one of the best early ideas was not spelled out in these RFCs, but only implicit in them: the consensus of structuring the protocols in layersso that one protocol could be based on another if desired, and programmers could write software that took advantage of whatever level of protocols best fit their needs.

We start with the bottom layer, the base. Steve Crocker wrote RFC 1 and Duvall wrote RFC 2. Together, these first two memos described basic streaming connections between hosts. They kept this layer simple: easy to define and easy to implement. Interactive terminal connections could be built on top of it (such as telnet), file transfer mechanisms (such as FTP) and other applications yet to be defined (such as email).

That was the plan, although it turned out to be more difficult than expected…

The handful of RFCs (Request for Comments) they circulated in early 1969 embodied the ideas about network protocols, but the work really began in earnest that September and October 1969, when the first IMPs arrived at UCLA and then at SRI. Just two were enough to start experimenting. Duvall at SRI and Charley Kline developed software that allowed a user of the machine to connect to the SRI machine.

On the afternoon of October 29, 1969, Charley tried unsuccessfully. After a quick fix to a small bug in the CRS software, a successful connection was achieved that same night. The software was adequate to connect UCLA to the SRI, but it was not general enough to connect all the machines that would eventually connect to the ARPANET. Had to keep working…

In February 1970there was a basic protocol of host to host communication that worked well enough to present it at the Joint Computer Conference that spring in Atlantic City. Within a few more months, the protocol was robust enough to shift attention to two application layer protocols, Telnet and FTP.

Instead of writing monolithic programs that run on every computer, as some of the team leaders had originally envisioned, they stuck to protocols building on each other to keep the system open and extensible. Designing Telnet and FTP to communicate over the host-to-host protocol ensured that they could be updated independently of the base system.

In October 1971, they were ready to put the ARPANET to the test.. They met at MIT for a complete test (they called it “the bake-off”), checking that every host could connect to every other host. It was a proud moment for the entire team, as well as a milestone that the Network Working Group had set for itself from the start.

Email begins to take shape

But there was still much to do. The network had grown to connect 23 hosts at 15 different sites.. A year later, in 1972 at a major communications conference in Washington, DC, the ARPANET was publicly demonstrated in a hotel ballroom. Visitors could sit in any of the terminals and connect to computers across the United States.

Year after year, the group continued to produce RFCs with comments, suggested changes, and possible extensions to the ARPANET and its protocols. Email was one of those early additions.. It started out as a specialized case of file transfer, but later evolved into a stand-alone protocol (Simple Mail Transfer Protocol, or SMTP, RFC 788, published in 1981). Email became the dominant use of the ARPANET.

Email also directly affected the team itself, as it allowed the group to circulate RFCs more quickly and to a much larger group of contributors. A virtuous cycle had begun: Each new function allowed programmers to create new ones more easily.

New Internet protocols are born

Protocol development flourished. TCP and IP protocols they replaced and greatly improved the host-to-host protocol and laid the foundations of the Internet. The RFC process led to the adoption of the Domain Name System (DNS, RFC 1035, published 1987), the Simple Network Management Protocol (SNMP, RFC 1157, 1990) and the Hypertext Transfer Protocol (HTTP, RFC 1945, 1996).

Over time, the development process evolved along with technology and the growing importance of the Internet in international communication and commerce. In 1979, Vint Cerf, then a program director at DARPA, created the Internet Configuration Control Council.which eventually gave rise to the Internet Engineering Task Force.

This working group continues the work originally done by the Network Working Group. Its members continue to discuss issues facing the network, modifications that might be needed to existing protocols, and new protocols that might be tools. And they continue to publish protocol specifications as documents labeled “Request for Comments” (RFC).

The core idea of ​​continuous improvement by consensus among a coalition of volunteers lives on in Internet culture.

Ideas about new protocols and changes to them are now spread through email lists dedicated to protocol-specific topics, known as working groups. Currently there are a hundred of these groups. When they meet at the triennial conferences, the organizers still don’t take a vote: They ask participants to comment on whether they agree with an idea, and then they discuss it in the room. Formal decisions are made after a subsequent exchange by email.

Draft protocol specifications are distributed as “internet drafts“, which are earmarked for discussion leading to an RFC. A recently started discussion on a new network software that allows quantum communication on the Internet, for example, is collected in an Internet draft of the RFC type.

And in an ironic twist, the specification of this or any new protocol will only appear in a request for comments after it has been approved for formal adoption and published. At that time, feedback is no longer solicited.

Although there is no specific date to celebrate the anniversary of the internetmany date it on October 29, 1969, at 10:30 p.m. We have related this little story of the beginnings and evolution of how the network of networks was created, you can get an idea of ​​how it all started and how it has been growing in consensus by many collaborators who maintain, manage and propose improvements so that the Internet continues to grow as technology grows.