Email: From Time Sharing To Mail Servers To The Cloud

With over 2.6 billion active users ad 4.6billion active accounts email has become a significant means of communication in the business, professional, academic, and personal worlds. Before email we had protocols that enabled us to send messages within small splinters of networks. Time Sharing systems like PLATO at the University of Champaign-Urbana, DTSS at Dartmouth College, BerkNet at the University of California Berkeley, and CTTS at MIT pioneered electronic communication. Private corporations like IBM launched VNET We could create files or send messages that were immediately transferred to other people. The universities that were experimenting with these messaging systems even used some of the words we use today. MIT’s CTSS used the MAIL program to send messages. Glenda Schroeder from there documented that messages would be placed into a MAIL BOX in 1965. She had already been instrumental in implementing the MULTICS shell that would later evolve into the Unix shell. Users dialed into the IBM 7094 mainframe and communicated within that walled garden with other users of the system. That was made possible after Tom Van Vleck and Noel Morris picked up her documentation and turned it into reality, writing the program in MAD or the Michigan Algorithm Decoder. But each system was different and mail didn’t flow between them. One issue was headers. These are the parts of a message that show what time the message was sent, who sent the message, a subject line, etc. Every team had different formats and requirements. The first attempt to formalize headers was made in RFC 561 by Abhay Bhutan and Ken Pogran from MIT, Jim White at Stanford, and Ray Tomlinson. Tomlinson was a programmer at Bolt Beranek and Newman. He defined the basic structure we use for email while working on a government-funded project at ARPANET (Advanced Research Projects Agency Network) in 1971. While there, he wrote a tool called CYPNET to send various objects over a network, then ported that into the SNDMSG program used to send messages between users of their TENEX system so people could send messages to other computers. The structure he chose was Username@Computername because it just made sense to send a message to a user on the computer that user was at. We still use that structure today, although the hostname transitioned to a fully qualified domain name a bit later. Given that he wanted to route messages between multiple computers, he had a keen interest in making sure other computers could interpret messages once received. The concept of instantaneous communication between computer scientists led to huge productivity gains and new, innovative ideas. People could reach out to others they had never met and get quick responses. No more walking to the other side of a college campus. Some even communicated primarily through the computers, taking terminals with them when they went on the road. Email was really the first killer app on the networks that would some day become the Internet. People quickly embraced this new technology. By 1975 almost 75% of the ARPANET traffic was electronic mails, which provided the idea to send these electronic mails to users on other computers and networks. Most universities that were getting mail only had one or two computers connected to ARPANET. Terminals were spread around campuses and even smaller microcomputers in places. This was before the DNS (Domain Name Service), so the name of the computer was still just a hostname from the hosts file and users needed to know which computer and what the correct username was to send mail to one another. Elizabeth “Jake” Feinler had been maintaining a hosts file to keep track of computers on the growing network when her employer Stanford was just starting the NIC, or Network Information Center. Once the Internet was formed that NIC would be the foundation or the InterNIC who managed the buying and selling of domain names once Paul Mockapetris formalized DNS in 1983. At this point, the number of computers was increasing and not all accepted mail on behalf of an organization. The Internet Service Providers (ISPs) began to connect people across the world to the Internet during the 1980s and for many people, electronic mail was the first practical application they used on the internet. This was made easier by the fact that the research community had already struggled with email standards and in 1981 had defined how servers sent mail to one another using the Simple Mail Transfer Protocol, or SMTP, in RFC 788, updated in 1982 with 821 and 822. Still, the computers at networks like CSNET received email and users dialed into those computers to read the email they stored. Remembering the name of the computer to send mail to was still difficult. By 1986 we also got the concept routing mail in RFC 974 from Craig Partridge. Here we got the first MX record. Those are DNS records that define the computer that received mail for a given domain name. So stanford.edu had a single computer that accepted mail for the university. These became known as mail servers. As the use of mail grew and reliance on mail increased, some had multiple mail servers for fault tolerance, for different departments, or to split the load between servers. We also saw some split various messaging roles up. A mail transfer agent, or MTA, sent mail between different servers. The received field in the header is stamped with the time the server acting as the MTA got an email. MTAs mostly used port 25 to transfer mail until SSL was introduced when port 587 started to be used for encrypted connections. Bandwidth and time on these computers was expensive. There was a cost to make a phone call to dial into a mail provider and providers often charged by the minute. So people also wanted to store their mail offline and then dial in to send messages and receive messages. Close enough to instant communication. So software was created to manage email storage, which we call a mail client or more formally a Mail User Agent, or MUA. This would be programs like Microsoft Outlook and Apple Mail today or even a web mail client as with Gmail. POP, or Post Office Protocol was written to facilitate that transaction in 1984. Receive mail over POP and send over SMTP. POP evolved over the years with POPv3 coming along in 1993. At this point we just needed a username and the domain name to send someone a message. But the number of messages was exploding. As were the needs. Let’s say a user needed to get their email on two different computers. POP mail needed to know to leave a copy of messages on servers. But then those messages all showed up as new on the next computer. So Mark Crispin developed IMAP, or Internet Message Access Protocol, in 1986, which left messages on the server and by IMAPv4 in the 1990s, was updated to the IMAPv4 we use today. Now mail clients had a few different options to use when accessing mail. Those previous RFCs focused on mail itself and the community could use tools like FTP to get files. But over time we also wanted to add attachments to emails so MIME, or Multipurpose Internet Mail Extensions became a standard with RFC 1341 in 1993. Those mail and RFC standards would evolve over the years to add better support for encapsulations and internationalization. With the more widespread use of electronic mail, the words were shortened and to email and became common in everyday conversations. With the necessary standards, the next few years saw a number of private vendors jump on the internet bandwagon and invest in providing mail to customers America Online added email in 1993, Echomail came along in 1994, Hotmail added advertisements to messages, launching in 1996, and Yahoo added mail in 1997. All of the portals added mail within a few years. The age of email kicked into high gear in the late 1990s, reaching 55 million users in 1997 and 400 million by 1999. During this time having an email address went from a luxury or curiosity to a societal and business expectation, like having a phone might be today. We also started to rely on digital contacts and calendars, and companies like HP released Personal Information Managers, or PIMs. Some companies wanted to sync those the same way they did email, so Microsoft Exchange was launched in 1996. That original concept went all the way back to PLATO in the 1960s with Dave Wooley’s PLATO NOTES and was Ray Ozzie’s inspiration when he wrote the commercial product that became Lotus Notes in 1989. Microsoft inspired Google who in turn inspired Microsoft to take Exchange to the cloud with Outlook.com. It hadn’t taken long after the concept of sending mail between computers was possible that we got spam. Then spam blockers and other technology to allow us to stay productive despite the thousands of messages from vendors desperately trying to sell us their goods through drip campaigns. We’ve even had legislation to limit the amount of spam, given that at one point over 9 out of 10 emails was spam. Diligent efforts have driven that number down to just shy of a third at this point. Email is now well over 40 years old and pretty much ubiquitous around the world. We’ve had other tools for instant messaging, messaging within every popular app, and group messaging products like bulletin boards online and now group instant messaging products like Slack and Microsoft Teams. We even have various forms of communication options integrated with one another. Like the ability to initiate a video call within Slack or Teams. Or the ability to toggle the Teams option when we send an invitation for a meeting in Outlook. Every few years there’s a new communication medium that some think will replace email. And yet email is as critical to our workflows today as it ever was.