RFC 32

SOME THOUGHTS ON SRI'S PROPOSED REAL TIME CLOCK ⏰✨

Title block tea 🫖💖

• Network Working Group 🌐
• Request for Comments: 32 🎀
• Author: Jerry Cole, UCLA 🧚
• Date: February 5, 1970 📅
• Obsoletes / Updates: the original doesn't list any 🙈
• Category / status: not stated in the original 💅

Re: NWG/RFC's #28 and 29 📚

Okay so bestie, a real time clock inside one or more of the network HOSTs is lowkey so desirable 😭🫶. It would let users measure message delays in a way that actually matters to them. Our current network measurement tools don't count any internal HOST delays 🙈, and those delays could be a big chunk of the total time a HOST-to-HOST message takes. ⏳ We might find that extending our "Trace" capabilities to include internal HOST delays is the perfect way to use a clock like this 🔥. To do that, you'd need a clock at both the source and the destination of the message 🧠, but those clocks don't have to be super accurate or synchronized with each other. 💅 Other tests, like measuring the absolute overall message delay from HOST A to HOST B, would need the two clocks to be synchronized ✨. That's a whole different vibe 😌.

Now for the specs ✨. A reasonable spec for the SRI real-time clock would have a resolution of about 1 msec. 🍓, an accuracy of about 1 part in 10E7 💜 (so two of these clocks could keep reasonable relative accuracy over many hours 🌙), and a range of about 24 hours ⭐. Honestly, a crystal controlled clock should easily hit all of this at a moderate cost 💰🦋. Iconic and very doable 👑.

The way the HOST actually reads the clock is also a big deal 👀. A 1 msec. resolution might mean the clock has to be entirely hardware 🔧 (as opposed to a core location that gets incremented on each clock pulse 🧮). That means the clock might need some pretty complicated interface circuitry 🫠. Not me getting stressed about wiring 💀.

At UCLA, we currently have two clocks on the Sigma 7 🖥️. One of them has a resolution of about 2 msec. 🌸, which could be usable for some internal HOST measurements 🎀. The catch is that it doesn't have the long term accuracy needed for the absolute measurements I mentioned earlier 😭. Main character problem, but for a different reason 💅.

[ This RFC was put into machine readable form for entry ] [ into the online RFC archives by Richard Ames 1/98 ] 🫶🌈