Method for providing time holdover in packet-based timing deployments
A packet network that includes the distribution of timing information (time and frequency) between server and client devices can continue to operate in a holdover mode even when the server loses its primary timing reference based on GNSS. This is achieved by populating the packet network with some client devices that also have access to the same timing reference. These devices are used in a reverse timing transfer mode to provide a hack-up reference to the server and thereby provide a graceful solution to the problem of loss of reference.
1. A method, comprising:
operating a packet-based communication network between a domain grand master clock, a first client clock coupled to the domain grand master clock and a second client clock coupled to the domain grand master clock,
providing time holdover in the packet based communication network when a time reference is not available at the domain grand master clock including:
synchronizing the second client clock to the domain grand master clock;
and
synchronizing the domain grand master clock to the first client clock when the time reference is available at the first client clock but not at the domain grand master clock; and
eliminating a need for the second client clock to switch over to an alternate grand master, reducing probability of a mobile network application outage,
wherein the time reference is available at the second client clock and
wherein the first client clock has the least noisy packet network with the domain grand master clock compared to the second client clock based on monitoring time delay variation as well as time delay.
2. The method of claim 1 , wherein operating the communication link between the domain grand master clock and the first client clock includes exchanging time-stamps associated with the time-of-arrival and time-of-departure of precision timing protocol packets.
3. The method of claim 2 , wherein operating includes calculating
Δ SM =t 4 −τ 3 (from an S-to-M packet)
Δ MS =τ 2 −t 1 (from a M-to-S packet)
wherein D MS is a transit delay from the domain grand master clock to the first client clock and D SM is a transit delay from the first client clock to the domain grand master clock,
wherein S denotes the first client clock and M denotes the domain grand master clock,
wherein t 1 is a time-of-departure of a packet transmitted by the domain grand master clock and
t 2 is a time-of-arrival of the packet transmitted by the domain grand master clock and
wherein t 3 is a time-of departure of a packet transmitted by the first client clock and t 4 is a time-of-arrival of the packet transmitted by the first client clock.
4. The method of claim 3 , wherein operating includes measuring asymmetry between the first client clock and the domain grand master clock.
5. The method of claim 3 , further comprising providing time holdover in the packet based communication network including synchronizing the domain grand master clock to the first client clock using D MS and D SM when the time reference is available at the first client clock but not at the domain grand master clock.
6. A computer readable medium, comprising non-transitory computer or machine readable program elements translatable for implementing the method of claim 1 .
7. An apparatus, comprising: a packet-based communication network including
a domain grand master clock;
a first client clock coupled to the domain grand master clock; and
a second client clock coupled to the domain grand master clock,
wherein time holdover is provided in the packet based communication network when a time reference is not available at one of the first client clock or the domain grand master clock including:
synchronizing the second client clock to the domain grand master clock;
synchronizing the first client clock to the domain grand master clock when the time reference is available at the domain grand master clock but not at the first client clock;
synchronizing the domain grand master clock to the first client clock using packet-based synchronization when the time reference is available at the first client clock but not at the domain grand master clock; and
eliminating a need for the second client clock to switch over to an alternate grand master, reducing probability of a mobile network application outage,
wherein the time reference is available at the second client clock and
wherein the first client clock has the least noisy packet network with the domain grand master clock compared to the second client clock based on monitoring time delay variation as well as time delay.
8. A mobile network comprising the apparatus of claim 7 .