3

Scott Manley, in his second video about communication satellites, focusing on the 1963 Telstar 1, mentioned at 10:22:

(Telstar 1) could be used for telephones it could carry multiple circuits, it carried data at something like 800 kilobits per second. They were able to exchange data between the computers. So the ground stations had these IBM 1620 computers that were needed to operate all the equipment and they were able to send the data from one computer to another, making it, well, kind of one of the earliest examples of satellite internet.

Telstar one was also used to synchronize clocks between Europe and the US. They had managed to get the clock synchronization down to 200 microseconds -- two milliseconds sorry 2,000 microseconds. But after Telstar they had got the clocks between these two continents synchronized to within one microsecond making the clock people very very happy.

Data exchange between computers is indeed not (as Manley points out) the same thing as "the internet", so I'd like to ask:

Question: What was the first satellite data link that can properly be called an internet connection?

15
  • 3
    ALOHAnet which pioneered, in a very practical demonstration, the networking protocol concept that later became CSMA and the foundation of Ethernet. Operational in 1971. Not satellite itself but medium-distance wireless (UHF frequencies). Predates the internet but was connected (via relays) to the ARPAnet. Linked article explains that the same techniques were used in an early satellite network. If not the first then certainly a contender.
    – davidbak
    Dec 2, 2023 at 23:30
  • 5
    DEC operated a US-UK satellite link as part of its internal network, from about 1983 onwards. This was 'production use', not experimental. It ran DECnet rather than TCP/IP, but DECnet had all the usual facilities (task-to-task, remote login, file transfer, mail, etc.).
    – dave
    Dec 3, 2023 at 0:52
  • 3
    How do you define internet connection? Since internet connections can be over phone, anything that could give you 300bps could have been used as an internet connection by someone, as soon as internet existed. Re soem of the comments: is ARPANET concidered Internet in this case?
    – UncleBod
    Dec 3, 2023 at 11:47
  • 3
    A link to the Video in question might a good idea. Not everyoen is aware of who "Scott Manley" is, and even if you are, it might be difficult to find the specific video.
    – UncleBod
    Dec 3, 2023 at 11:51
  • 4
    @SolomonSlow - I disagree. It's inter-networking if it connects a couple of networks. And since we have local-area networks, connecting two of those up, even if they belong to the same organization, is inter-networking,
    – dave
    Dec 4, 2023 at 3:27

2 Answers 2

8

What was the first satellite data link that can properly be called an internet connection?

I'd say the first dial up across the pond that by chance happened to end up on a satellite connection:)

A less random candidate might be the SATNET demonstration of late 1977. Here satellite as well as mobile data links were used to transport TCP controlled packets between the California (US), England and Norway, using Intelsat IV-A and ground stations in West Virginia, US, Goonhilly, England and Tanum, Sweden.

enter image description here

(Picture taken from Wikipedia)

While the satellite link to Tanum was already used since 1973 for an ARPANET connection to NORSAR(*1), but of course with ARPANET's own protocol. After all, it was the time of great plurality: Many networks wit even more protocols:))

SATNET was the very reason the Internet Protocol was developed by Vint Cerf and others. SATNET, envisioned by Bob Kahn, was a packet network like ARPANET, but using its own protocols. It was developed with a focus on a single satellite connections shared between stations. By the mid 1970s it connected ground stations in Germany, Italy, Norway, UK and US. Those data channels were in parallel to the ARPANET ones.

Kahn thought about bridging between those networks by implementing an _Inter_net protocol allowing to maintain and route connections across different networks. Vint Cerf joined and combining Kahn's ideas with principles from the French CYCLADES network, which unlike ARPANET or SATNET was already designed for inter-networking. The result was TCP/IP and above setup of 1977, which might be considered not only the first satellite using Internet but also the first Internet at all :)


It's Complicated:

It feels as if this question is based on a false dichotomy of 'The Internet' vs. 'a' data link. If not a trichotomy.

The internet as we know it today is collection of applications using on the internet protocol suite, which in turn uses data links to transfer information. Neither of this defines how many computers are involved (*2),nor does it define what functionality must exist to call data link use 'Internet'. Thus any question using this dichotomy may be invalid in itself.

In addition, the question itself already illuminates the issue of fuzzy definition even further:

What was the first satellite data link that can properly be called an internet connection?

(Emphasis mine)

It's That 'Proper' Part Highlightening the Whole Issue

Internet is not only a very vague definition (*3) but also one that changed over time - not at least by gradual development of the protocols that build the internet.

Putting commercial/manufacturer specific applications like Facebook, Twitter or Instagram aside, It's safe to state that everything Joe Average sees is HTTP based - a protocol as recent as 1991 (*4).

Can There Be Internet Proper Without HTTP i.e. before 1991?

Older gents like yours truly will of course object and point to SMTP and NNTP being all they need. Except, that only pushes the date back at most to 1980 when SMTP application level protocols were moved to IP .Which also moves the question to

Can There Be An Internet Proper Without Application Protocols i.e. before 1980?

1980 is not only the year of many applications moving toward IP but also when the IP suite was extended with UDP. Which kicks the can further down the road as 'The Internet Protocol' and in turn TCP was only developed in 1974. Thus:

Can There Be An Internet Proper Without Internet Protocol itself i.e. before 1974?

Some may say yes, as the ARPANET, cited as the start of the internet (*5), started operations in 1967

So What?

Don't ask me. it all depends on your value of 'proper' Internet. One may be eager to find something spectacular early and twist as many adjectives as needed, or come down to a rather sober conclusion that satellite based internet is a nice thing to support remote operation, but otherwise was already outdated when the Internet became the Internet we know.

All of that still based on a

Regarding Mr.Manley point I'd say it's neither right nor wrong. It's a valid description for today's viewers about an achievement at a time when the internet we know was still just some fluctuations in the primordial soup it would some day evolve from.


*1 - Which in turn was already connected to ARPANET as early as 1971.

*2 - Well, commonly at least two would be nice, but all of those applications work as fine using localhost :))

*3 - For some people and/or even whole countries Facebook or Tick-Tock is the internet.

*4 - Though, he might name whatever browser he uses.

*5 - More rational, the ARPANET was just one of many networking attempts during the 60s and 70s. neither the first not the definitive one.

9
  • 1
    @JonCuster clarified that part, hope it helps.
    – Raffzahn
    Dec 3, 2023 at 22:27
  • 1
    If the entire network is running IP, is there any 'inter-networking' at all? ;-)
    – dave
    Dec 4, 2023 at 3:30
  • @another-dave - I get your point, but a modern interpretation might be that the internet starts where boundary routing protocols live. I think "inter networking" started earlier than boundary protocols - weren't there ARPAnet/DECnet bridges before those, for example - but I think there's more to "internetworking" than saying that we don't have it anymore because IP makes the entire world one network.
    – davidbak
    Dec 4, 2023 at 22:09
  • @davidbak Gateways aren't really inter-networking. They just translate between specific networks using very specific rules and assumptions only fitting those two networks. Any message that needs to pass thru those gateways have to be made according to those rules, carrying enough information to satisfy both sides. Now add a third network, which complicates this even more, adding many more restrictions. Long story short: Gateways are not inter-networking but point to point connections. Inter-networking means that messages are formed and exchanged independent of the network(s) used inbetween.
    – Raffzahn
    Dec 4, 2023 at 22:44
  • 1
    Personally I've never heard that specific definition of internet/internetworking. And, frankly, I don't think it's correct. But I don't have the time right now to look further. So ... no conclusion at this time.
    – davidbak
    Dec 5, 2023 at 2:48
5

The main question is "define Internet": When we're talking about the commercial Internet, that really only took off in the late 1980s, early 1990, this coincides with the development and commercial introduction of satellite TV which really made transponder rental and thus dedicated satellite network connections commercially viable. Satellite datacomms is, astonishingly, much older than "the Internet" (whatever that might be...)

If you go earlier, you can find the necessary technical capabilities fulfilled as early as 1965, starting with Intelsat F1, which had two C-Band transponders that could provide (serial) data connections. Thus, data connections provided via satellite pre-date the introduction of "Internet protocols" (as standards) by about 15 years.

"Traditional" satcom is really not much of a technical challenge for an internet connection - Because the satellite transponder is in principle nothing but a big "mirror in space", it behaves like an (admittedly very long) piece of wire, typically at T1 (1.5Mbps) or E1 (2Mmbps) data rate, to the connection. TCP connections simply increase their window size to adapt to the longer round-trips. UDP and interactive connections, however, tend to suffer from the roughly 3-500ms delay on the data path.

The timeframe mentioned above coincides with the introduction of satellite-based TV, which did provide satellite transponder capacity that was "affordable" for the first time and enabled commercially viable satellite internet (with all the downsides of comms over geo-stationary satellites).

This "old-style" technology is in nearly no way comparable with contemporary satellite Internet like StarLink - That works with low-earth-orbit satellites that are interconnected through optical laser links (thus, the sender and receiver do not need to be both in the same coverage area of one single satellite, and the round-trip-delay is much lower because of the low orbits of the satellites). Thus, Starlink is a real network compared to classical satcom, which typically provides a point-to-point connection.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .