Index Journeys
RFC 29 UNKNOWN ARPANET · TIMING

Response to RFC 28

The IMP uses two clocks: a fine hardware clock for internal timing and a coarser software clock for measurements.

On rfc-editor.org ↗
PUBLISHED
January 1970
AUTHORS
R.E. Kahn
STREAM
Legacy
HARDWARE CLOCK RESOLUTION
100 microseconds
SOFTWARE CLOCK RESOLUTION
25.6 milliseconds
DOI
10.17487/RFC0029

RFC text

Original text · RFC Editor View on rfc-editor.org ↗ 
Network Working Group                                   Robert Kahn
RFC-29                                                  BBN
                                                        19 January 1970





This note is in response to Bill English's Request for Comments:  28.

A "millisecond" clock should be satisfactory for most network measurements.
Round-trip message transit times typically should be at least on the
order of tens to hundreds of milliseconds.  The IMP contains a 16-bit
hardware clock which is incremented every 100 microseconds to allow for
timing of internal events within the IMP, as for example, during tracing.
However, most measurements are made using a 25.6 ms. software clock.














DISTRIBUTION:

A. Bhushan, MIT
S. Carr. Utah
G. Cole, SDS
S. Crocker, UCLA
K. Fry, MITRE
J. Heafner, RAND
B. Kahn, BBN
T. O'Sullivan, Raytheon
L. Roberts, ARPA
P. Rovner, LL
R. Stoughton, UCSB

AI ELI5

On this page
PROBLEMRFC 28 raised questions about time standards for network measurements.
MECHANISMThe IMP contains a 16-bit hardware clock incremented every 100 microseconds for internal timing and tracing. A separate 25.6 millisecond software clock is used for most network measurements.
TIMING SPECIFICATIONThis two-tier timing approach resolved the timing requirements for early ARPANET operations.

§1Network Timing

RFC 29 responds to Bill English's RFC 28 questions about time standards. A millisecond clock is satisfactory for most network measurements, since round-trip message transit times are typically at least tens to hundreds of milliseconds.

Measuring how long messages take does not need very precise timing—a millisecond is good enough.

§2Hardware Clock

The IMP contains a 16-bit hardware clock that is incremented every 100 microseconds. This fine-grained clock allows for precise timing of internal events within the IMP, such as during tracing and internal event sequencing. The hardware clock provides microsecond-level granularity for debugging and understanding the IMP's internal behavior.

The IMP's fast clock ticks every 100 millionths of a second, which is fast enough to time internal events.

§3Software Clock for Measurements

Most network measurements use a coarser 25.6 millisecond software clock. This granularity is adequate because round-trip message transit times are typically at least tens to hundreds of milliseconds. Using a software clock for measurements reduces the processing burden on the IMP compared to using the hardware clock for every measurement.

A slow clock that ticks every 25 milliseconds is accurate enough for measuring message delays across the network.