System for establishing and maintaining a clock reference indicating one-way latency in a data network

• US 8,792,380 B2
• Filed: 08/24/2012
• Issued: 07/29/2014
• Est. Priority Date: 08/24/2012
• Status: Active Grant

• Current Assignee: Accedian Networks Inc. (Cisco Systems, Inc.)
• Original Assignee: Accedian Networks Inc. (Cisco Systems, Inc.)
• Inventors: Hansson, Kjell, Hagsand, Olof
• Primary Examiner(s): Lee, Andrew
• Application Number:
  US13/593,888
• Publication Number:
  US 20140056318A1
• Time in Patent Office: 704 Days
• Field of Search: 370/252, 370/503, 370/509
• US Class Current: 370/252
• CPC Class Codes: H04J 3/0658 Clock or time synchronisation among packet nodes
• CPC Class Codes: H04J 3/0667 Bidirectional timestamps, e.g. NTP or PTP for compensation of clock drift and for compensation of propagation delays arrangements for monitoring round trip delays in packet switching networks H04L43/0864
• CPC Class Codes: H04L 1/0018 based on latency requirement
• CPC Class Codes: H04L 43/0858 One way delays
• CPC Class Codes: H04L 43/0864 Round trip delays
• CPC Class Codes: H04L 69/22 Parsing or analysis of headers
• CPC Class Codes: H04L 7/0012 by comparing receiver clock with transmitter clock

Abstract

A method for indicating one-way latency in a data network, with continuous clock synchronization, between first and second node having clocks that are not synchronized with each other includes a continuous synchronisation session and a measurement session. The method repetitively sends predetermined synchronization messages from the first node to the second node and from the second node to the first node, calculates a round trip time for each message at the first node, updates a synchronization point if the calculated round trip time is smaller than a previously calculated round trip time, stores the updated synchronization points of a synchronization window, and calculates a virtual clock from the updated synchronization points of the synchronization window. The measurement session collects multiple measurements of one-way latency between the first and second nodes using the virtual clock, and generates a latency profile by interpolating the multiple measurements.

Claims

1. A method for indicating one-way latency in a data network, with continuous clock synchronization, between first and second nodes having clocks that are not synchronized with each other, comprising:

executing a continuous synchronisation session by repetitively

sending predetermined synchronization messages from the first node to the second node and from the second node to the first node,

calculating a round trip time for each message at the first node,

updating a synchronization point if the calculated round trip time is smaller than a previously calculated round trip time,

storing the updated synchronization points of a synchronization window, and

calculating a virtual clock from the updated synchronization points of said synchronization window, and

executing a measurement session to determine the one-way latency of said network between said first node and said second node by

collecting multiple measurements of one-way latency between said first and second nodes using the virtual clock, and

generating a latency profile by interpolating said multiple measurements.

2. A method according to claim 1 which includes a plurality of said second nodes.

3. A method according to claim 1 in which said a round trip time RT is calculated according to RT=T4−

T1−

(T3−

T2),

where T1 is the sending time from the first node,

T2 is the receiving time at the second node,

T3 is the sending time at the second node, and

T4 is the receiving time at the first node.

4. A method according to claim 1 in which said updating of a synchronization point includes

comparing the calculated round trip time with the stored minimum round trip time for the current synchronization point and, if less,

updating the stored minimum round trip time with the calculated round trip time for said current synchronization point, and

calculating an offset CDIFF according to CDIFF=(T2−

T1+T3−

T4)/2, where

T1 is the sending time at the first node,

T2 is the receiving time at the second node,

T3 is the sending time at the second node, and

T4 is the receiving time at the first node, and

setting the absolute clock (CABS) value for said current synchronization point to T4.

5. A method according to claim 4 in which said storing of a new synchronization point in a synchronization window includes

storing the minimum round trip time value (RTTMIN),

storing the calculated offset (CDIFF), and

storing the absolute clock value (CABS).

6. A method according to claim 1 in which said calculating of a virtual clock includes using the sum of least square method, comprising

calculating the mean value for the absolute clock value (CABSmean) and the mean value for the offset (CDIFFmean) for each synchronization point in said synchronization window;

calculating the sum of differences above the slope of said synchronization window for each synchronization point (ABOVE) as the sum of (CABS−

CABSmean)×

CDIFF;

calculating the sum of differences below the slope of said synchronization window for each synchronization point (BELOW) as the sum of (CABS−

CABSmean)×

(CABS−

CABSmean);

calculating a skew value (SKEW) as said sum (ABOVE) divided by said sum (BELOW);

calculating a new offset CDIFF as CDIFFmean−

(SKEW×

CABSmean);

setting the new absolute clock value CABS for said virtual clock to CABSmean, and;

computing the clock difference CDIFF from time zero to the current reference point by adding the product of the absolute clock value and the skew to the previously calculated value of the offset CDIFF.

7. A method according to claim 1 which includes, for each of a plurality of measurement messages received at the second node,

sending said predetermined synchronization message N times from the first node to the second node chronologically equidistantly,

receiving said predetermined synchronization message at the second node,

storing the sending time value for the first node,

storing the receiving time value for the second node, and

storing a validity indicator for each valid message.

8. A method according to claim 7 which includes measuring the overhead for making measurements at the first node.

9. A method according to claim 7 which includes measuring the overhead for making measurements at the second node.

10. A method according to claim 1 in which said interpolating comprises calculating the one-way latency in the data network between a first node and a second node according to:

T2−

T1+CDIFF.

11. A method according to claim 1 in which said interpolating comprises calculating the one-way latency in the data network between a first node and a second node, for each of a plurality of messages received at the second node, by calculating the one-way latency in the data network between a first node and a second node according to:

T2−

T1+CDIFF.

12. A method according to claim 10 which includes

an overhead term corresponding to the overhead for making measurements at least one of the first and second nodes, and

calculating the one-way latency in the data network between a first node and a second node according to:

T2−

T1+CDIFF−

the sending overhead term−

the receiving overhead term.

13. A method according to claim 11 in which said interpolating comprises calculating the one-way latency in the data network between a first node and a second node, for each of a plurality of messages received at the second node, by calculating the one-way latency in the data network between a first node and a second node according to:

T2−

T1+CDIFF−

the sending overhead term−

the receiving overhead term.