Time Sync Network Limits: Status, Challenges Joint IEEE-SA and ITU Workshop on Ethernet Time Sync Network Limits: Status, Challenges Stefano Ruffini, Ericsson Q13/15 AR Geneva, Switzerland, 13 July 2013

Contents Introduction on G.8271 and G.8271.1 Definition of Time sync Network Limits Challenges for an operator Next Steps Geneva, Switzerland, 13 July 2013

Time Sync: Q13/15 Recommendations 2013-04-14 Analysis of Time/phase synchronization in Q13/15: G.8260 (definitions related to timing over packet networks) G.827x series Frequency Phase/Time General/Network Requirements Architecture and Methods PTP Profile Clocks G.8261 G.8261.1 G.8264 G.8265 G.8265.1 G.8262 G.8263 G.8271 G.8271.1 G.8275 G.8275.1, G.8275.2 G.8272 G.8273,.1,.2,.3 G.8271.2 ITU-T Q13/15 focusing on this topic after December 2011 (when most of frequency sync aspects were completed) The following main aspects are addressed Network Requirements Architecture PTP Profiles Clocks The work is tentatively planned to be completed in the 2013/2014 time frame Several aspects also involving other SG15 Questions, e.g.: Time sync Interfaces Time sync over access technologies Time sync over OTN G.8271 scope Time and phase synchronization aspects in packet networks Target applications Methods to distribute the reference timing signals It also specifies the relevant time and phase synchronization interfaces and related performance. Physical characteristics have been into G.703 (Q11/15) G.8271 is the first document of the G.827x series to be released (Published in 02/2012) Amendment planned for July 2013 (additional details and alignments with G.8271.1) G.8271.1Scope maximum network limits of phase and time error that shall not be exceeded. minimum equipment tolerance to phase and time error that shall be provided at the boundary of these packet networks at phase and time synchronization interfaces. Related Information (HRM, simulation assumptions, etc.) Consented in July 2013 ? (TD xxx) Details on Simulations in G.Supp Planned for 2013 Geneva, Switzerland, 13 July 2013 © Ericsson AB 2013

Target Applications Level of Accuracy 2013-04-14 Target Applications Level of Accuracy Time Error Requirement (with respect to an ideal reference) Typical Applications 1 500 ms Billing, Alarms 2 100 ms IP Delay monitoring 3 5 ms LTE TDD (cell >3km) 4 1.5 ms UTRA-TDD, LTE-TDD (cell  3Km) Wimax-TDD (some configurations) 5 1 ms Wimax-TDD (some configurations) 6 < x ns (x ffs) Location Based services and some LTE-A features (Under Study) Target application is 1.5 us Geneva, Switzerland, 13 July 2013 4 © Ericsson AB 2013

Time sync Network Limits Aspects to be addressed when defining the Network Limits Reference network (HRM) for the simulations Metrics Network Limits Components (Constant and Dynamic Time Error) Failure conditions Network Rearrangements Time Sync Holdover Geneva, Switzerland, 2 13 July 2013

Noise (Time Error) Budgeting Analysis 2013-04-14 Noise (Time Error) Budgeting Analysis Common Time Reference (e.g. GPS time) N TEA TEB Same limit applicable to R3 and R4 (limits in R4 applicable only in case of External Packet Slave Clock) TEC Network Time Reference (e.g. GNSS Engine) R4 R5 R1 R2 R3 Simulation Reference Model: chain of T-GM, 10 T-BCs, T-TSC with and without SyncE support PRTC Packet Master (T-GM) Packet Network Packet Slave Clock (T-TSC) End Application Time Clock T-BC: Telecom Boundary Clock PRTC: Primary Reference Time Clock T-TSC: Telecom Time Slave Clock T-GM: Telecom Grandmaster Typical Target Requirements TED < 1.5 ms (LTE TDD, TD-SCDMA) Geneva, Switzerland, 13 July 2013 © Ericsson AB 2013

Rearrangements and Holdover 2013-04-14 The full analysis of time error budgeting includes also allocating a suitable budget to a term modelling Holdover and Rearrangements Time Sync Holdover Scenarios PTP traceability is lost and and the End Application or the PRTC enters holdover using SyncE or a local oscillator PTP Master Rearrangement Scenarios PTP traceability to the primary master is lost; the T-BC or the End Application switches to a backup PTP reference TE (t) t |TE| Holdover-Rearr. period TEHO or TEREA budget 1.5 us Failure in the sync network TEHO applicable to the network (End Application continues to be locked to the external reference) TEREA applicable to the End Application (End Application enters holdover) Geneva, Switzerland, 13 July 2013 © Ericsson AB 2013

Sync Standards Coordination 2013-05-14 MAX |TE| based Limits The Constant Time Error measurement was initially proposed as could be easily correlate to the error sources (e.g. Asymmetries), however Complex estimator (see G.8260) Different values at different times (e.g. due to temperature variation) Max |TE| has then been selected : The measurement might need to be done on pre-filtered signal (e.g. emulating the End Application filter, i.e. 0.1 Hz). This is still under study. End Application T-TSC TED D TE(t) PTP 1 PPS C Max|TE| max|TE’| Test Equipment 1500 ns Max |TEC (t)| = max|TE’| + TEREA + TEEA < TED Geneva, Switzerland, 13 July 2013 1100 ns 400 ns

Time Error Budgeting 1.1 us Network Limit (max |TE|) 2013-04-14 Time Error Budgeting 1500 ns Budgeting Example (10 hops) Dynamic Error (dTE (t)) simulations performed using HRM with SyncE support It looks feasible to control the max |TE| in the 200 ns range Constant Time Error (cTE) Constant Time Error per node: 50 ns PRTC (see G.8272): 100 ns End Application: 150 ns Rearrangements: 250 ns (one of the main examples) Remaining budget to Link Asymmetries (250 ns) Holdover PTP Rearrangements 250ns End Application 150 ns 1.1 us Network Limit (max |TE|) Dynamic Noise accumulation 200 ns BC Internal Errors (Constant) 550 ns Link Asymmetries 250 ns PRTC 100 ns Geneva, Switzerland,13 July 2013 © Ericsson AB 2013

Stability Requirements Additional requirement on stability of the timing signal is needed and is under study Applicable to the dynamic component (d(t)) In terms of MTIE and TDEV Possible Jitter requirements Important for End Application Tolerance Geneva, Switzerland,13 July 2013

Challenges for an operator Distribution of accurate time synchronization creates new challenges for an operator Operation of the network Handling of asymmetries (at set up and during operation) Planning of proper Redundancy (e.g. Time sync Holdover is only available for limited periods (minutes instead of days). Exceeding the limits can cause service degradation New testing procedures Network performance and Node performance requires new methods and test equipment Some aspect still under definition (e.g. G.8273.x) Geneva, Switzerland, 2 13 July 2013

Sources of Asymmetries Frequency and time/phase synchronization in Packet Networks 3/28/20172011-10-19 Sources of Asymmetries Different Fiber Lengths in the forward and reverse direction Main problem: DCF (Dispersion Compensated Fiber) Different Wavelengths used on the forward and reverse direction Asymmetries added by specific access and transport technologies GPON VDSL2 Microwave OTN Additional sources of asymmetries in case of partial support : Different load in the forward and reverse direction Use of interfaces with different speed Different paths in Packet networks (mainly relevant in case of partial support) Traffic Engineering rules in order to define always the same path for the forward and reverse directions Geneva, Switzerland,13 July 2013 © Ericsson AB 2013

Next Steps Work is not completed Dynamic components in terms of MTIE and TDEV; Jitter? Testing methods (G.8273 provides initial information) Partial Timing support Geneva, Switzerland, 13 July 2013

Partial Timing Support HRM for G.8271.2 Need to define new metrics (e.g. 2-ways FPP) Geneva, Switzerland,13 July 2013

Summary G.8271.1 consented this week: Max |TE| Time sync limits are available The delivery of accurate time sync presents some challenges for an operator Asymmetry calibration Handling of failures in the network Still some important topics need to be completed Stability requirements Partial timing support (G.8271.2) Geneva, Switzerland, 13 July 2013

Back Up Geneva, Switzerland, 13 July 2013

Time Synchronization via PTP Frequency and time/phase synchronization in Packet Networks 3/28/20172011-10-19 Time Synchronization via PTP The basic principle is to distribute Time sync reference by means of two-way time stamps exchange Symmetric paths are required: Basic assumption: t2 – t1 = t4 – t3 Any asymmetry will contribute with half of that to the error in the time offset calculation (e.g. 3 ms asymmetry would exceed the target requirement of 1.5 ms) M S t1 Time Offset= t2 – t1 – Mean path delay Mean path delay = ((t2 – t1) + (t4 – t3)) /2 t2 t3 t4 Geneva, Switzerland,13 July 2013 © Ericsson AB 2013

Metrics TE (t) Max |TE| t 2013-04-14 Metrics Main Focus is Max Absolute Time Error (Max |TE|) (based on requirements on the radio interface for mobile applications) Measurement details need further discussion Stability aspects also important MTIE and TDEV Related to End Application tolerance Same Limits in Reference point C or D ! Same limits irrespectively if time sync is distributed with SyncE support or not ? TE (t) Max |TE| t Geneva, Switzerland, 13 July 2013 © Ericsson AB 2013

Measurement of Newtork Limits at ref. Point C Frequency and time/phase synchronization in Packet Networks Measurement of Newtork Limits at ref. Point C 3/28/20172011-10-19 Option a C End Application ... T-TSC PRTC T-GM T-BC T-BC 1pps Direct comparison of PRTC Time with 1 PPS PRTC Test Equipment ... PRTC T-GM T-BC C Monitoring method (e.g. tapped) PTP Monitor Test Equipment T-TSC End Application X TE = TM1 - T1 – X = TM4 – T4 -X Sync timestamp Monitor timestamp of Sync Delay Resp (t4 timestamp) Monitor timestamp of Delay Request Option b In alternative the 4 timestamps could be used: TE = (TM2 – T1 – T4 + TM3)/2 Geneva, Switzerland,13 July 2013 © Ericsson AB 2013

Frequency and time/phase synchronization in Packet Networks 3/28/20172011-10-19 Measurement of Newtork Limits at ref. Point C Option c from the two-way PTP flow via an active measurement probe (e.g. prior to the start of the service, or connecting the active monitor to a dedicated port of the T-BC). C End Application ... T-TSC PRTC T-GM T-BC T-BC Monitoring method (active probe) PRTC Test Equipment PTP Probe Geneva, Switzerland,13 July 2013 © Ericsson AB 2013

Example of Time Error Accumulation 2013-04-14 Accumulation of maximum absolute time error over a chain of boundary clocks for different values of asymmetry bias. The physical layer assist involves SEC/EEC chain with bandwidth 10Hz. 40 ns Time Error due to asymmetry per hop (Constant Time Error) Source: WD25 (Anue), York, September 2011 no asymmetry (only Dynamic component) Geneva, Switzerland,13 July 2013 v = max asymmetry per hop © Ericsson AB 2013