Internet dies if GPS dies?

[Dave Emery]

On Sun, Apr 20, 2003 at 09:35:43PM -0700, Major Variola (ret) wrote:
> [If GPS dies] "Internet activity would slow to a crawl, because many
>   backbone operators rely on precise GPS time stamps to route data. "
> http://www.wired.com/wired/archive/11.05/start.html?pg=6
> 
> Sounds like bullshit to me, data clocks should be able to run without
> being fully synched externally, constantly.
> Maybe very occasional minor glitches at boundaries of clock domains.
> 

	Much of the current telcom plant from CDMA cellphones to many
high capacity fiber muxes, switches and rings depend on GPS timing and
operate with a decentralized timing hierarchy with each major node
having its own primary reference locked to GPS time.   And CDMA
cellsites in particular cannot perform proper handoffs without GPS
accuracy timing.  Even such technology as low tech as police radio
trunking systems and pagers now depend on GPS for accurately
synchronizing multiple ("simulcast") transmitters so they don't interfere
with each other.

	It is, however, true that most of these facilities use
accurate rubidium or sometimes even cesium clocks as backup for
GPS outages and even CDMA cellsites often have high precision ovenized
oscillator frequency standards that will meet the timing drift specs
for up to a couple of days with no valid GPS signal.

	And some network facilities can fall back to locking their
timing to other facilities (eg the data coming in from a fiber)  rather
than GPS - it really depends on how much care and thought went into the
system design and whether any of the backup timing modes are tested
regularly.

	But yes, GPS time and frequency is VERY widely used, and has
revolutionized the timing and synchronizing architectures of many
kinds of telcom networks and systems.   It ain't 1975 when all timing
and frequency for the whole CONUS radiated out of a master oscillator
in an underground bunker in Hillsboro Mo, and essentially everything
was slaved to it by layers of PLLs and local disciplined oscillators.

	And having a GPS broadcast time base has made life much easier
for system designers - time of day accurate to around 25 ns is now
available for less than $1K anywhere you want it with very high
reliability as is frequency accurate to parts in 10^13.   Back in the old
days these were unheard of numbers... with maybe 10 ms timing accuracy
being considered good.

	And yes, all of this is to some degree or another - sometimes
quite severely - vulnerable to a long term GPS outage of hours or
days or longer.   Virtually all of these GPS timing systems are designed
to stand short GPS outages, as they can happen due to sky coverage,
satellite failures, or routine maintenance.

	And yes, malicious jamming is a real threat.

	What else is new ?

-- 
	Dave Emery N1PRE,  die at die.com  DIE Consulting, Weston, Mass 02493
PGP fingerprint 1024D/8074C7AB 094B E58B 4F74 00C2 D8A6 B987 FB7D F8BA 8074 C7AB