Transmission - Synchronization Design

Embed Size (px)

Citation preview

  • 8/3/2019 Transmission - Synchronization Design

    1/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    Synchronization Design

    1 of 14

  • 8/3/2019 Transmission - Synchronization Design

    2/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    1. Overview

    Synchronization is a fundamental component of an optimally performing network. Synchronizationnetworks provide timing signals to all synchronization network elements at each node in a digital network.These timing signals are traceable to a highly accurate PRS (Primary Reference Source) clock ofStratum 1 quality with frequency accuracy of 10-11 or better. The aim is to ensure that all outgoingtransmissions from a digital network node have the same average frequency. Buffer elements are usedat important transmission interfaces to absorb differences between the average local frequency and theactual short-term frequency of incoming signals, which may be affected by phase wander and jitteraccumulated along the transmission paths.

    The very nature of digital communications is based on clocking signals. All digital signals require someform of clocking. Sampling, multiplexing, accessing and transportation of digital data are all dependent on

    changes based on prescribed time intervals. Digital switches, digital cross-connect, multiplexers anddigital transport systems are all linked for the sole purpose of reliable communications. All of these piecesmust mesh with respect to time in order to get the right information sampled, switched, transported andreceived.

    Network synchronization is a method of achieving:

    A common clock rate for the elements making up the network and;

    Ensuring that transmission impairments impacting synchronization between nodes be kept to aminimum.

    In addition to synchronizing the clock rates at the transmit and receive ends, the links connecting both

    ends must be designed to minimize timing impairments such as jitter and frequency wander. The qualityof the transmission facilities for transporting traffic or synchronization data is dependent on the degree ofalignment between the transmit and receive clocks. The higher the degree of alignment the better theerror performance. As the degree of alignment degrades, the rates of operation at the transmit andreceive nodes start to differ. The end result of this difference in rates is a loss or repetition of bits or bytesof information. This event is defined a slip.

    Digital communications relies on the use of data buffers to account for differences in clock rates betweenthe transmitted and received data. As the difference increases, the amount of buffer capacity consumedgrows. If the difference is too great for the buffers, then an overflow occurs resulting in a slip condition.

    The impact of slip degradation on different services are shown in the following table:

    Service Observed effectVoice Occasional audible clicks

    Voice band data Serious degradation, with some modems needing up to 6 secondsto recover from a slip

    Fax Each slip causes compression or blank lines

    Encrypted voice Serious degradation requiring retransmission of the encryption key

    Compressed video Serious degradation resulting in freeze frames or missing lines

    Packet data Loss of packets, resulting in degraded throughput andretransmission delays

    Table 1: Impact of slips

    2 of 14

  • 8/3/2019 Transmission - Synchronization Design

    3/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    In the PDH environment, the classical mechanism for passing synchronization through a telecom networkis to use the TS0 (Time Slot 0). In order to overcome slip problems in the PDH networks, SDH employs acontainer structure to move the traffic through the network. This considerably reduces the impact of poorsynchronization on the efficiency of the transmission network and correspondingly improving thethroughput of applications. As the containers realign themselves within the transmission layer, they injectphase hits onto the synchronization layer. The result is that TS0 in an E1 delivered over an SDH networkwould be susceptible to phase hits if the network is to become badly behaved and experience pointeradjustments. Any applications dependent on a clean, stable wander-free TS0 is potentially at risk.

    A synchronization problem exists in two distinct forms:

    Non-systematic, i.e. spontaneous and unpredictable: As the virtual container is moved around toaccommodate anomalies in transmission, TS0 experiences either VC-4 or VC-12 phase hits. VC-

    4 pointers occur if the virtual container is in the higher level and VC-12 pointers occur if thevirtual container is in the lower level. VC-4 pointer adjustments have less of an impact thanVC-12, being 160 nanoseconds rather than 3.57 microseconds.

    Systematic (i.e. regular and continuous) phase transients are created within the variousequipment elements in a telecom network by either the phase locked loop behaviour of the clockrecovery circuit or the effects of buffer correction activity.

    Both types of noise will be a problem where the connected application is expecting to see good qualitysynchronization on the E1.

    What is causing switching phenomenon? In legacy PDH networks, long term frequency drift or very long

    wander build up within a network must be compensated for at the access point. If not, it will causeproblems by filling up the buffer and creating a slip.

    Packetization delay is the delay caused by the necessity to fill a packet before it is transmitted and isparticularly evident with compressed voice calls associated with GSM traffic. Buffering delay results fromthe need to maintain a real time delivery of voice traffic across the network. The traffic must be brokendown into packets in order to be transmitted. These packets must then be reassembled in order toreproduce (emulate) the original voice call. This delay can often be considerable and if it is not carefullycontrolled can cause problems with CBR traffic such as voice over ATM. However the controlmechanism, i.e. the buffer imposes its own effect on the E1, experienced as the frequency switching.

    2. Synchronization Elements and Standards

    The digital switch is made up of timed elements which may consists of cross connects, multiplexers andgateways. The deployment of these devices is based on the need to transport and route traffic amongst anetwork of telecommunications nodes. The network design for these elements is a function of access,traffic patterns, carrier boundaries and circuit performance criteria.

    The circuit performance criteria is predicated on a maximum allowable level of jitter and wander. Theamount of network jitter and wander present is dependent on the type of clocks used and the criteria fornetwork design. Choosing the right type of clock is the first step in providing a solid infrastructure for thetimed network elements.

    Master-slave synchronization uses a hierarchy of clocks in which each level of the hierarchy issynchronized with reference to a higher level, the highest level being the PRC.

    3 of 14

  • 8/3/2019 Transmission - Synchronization Design

    4/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    The hierarchical levels are shown below:

    The PRC (G.811) may be realized as an autonomous clock operating independently of any othersources, or as a non-autonomous clock disciplined by a UTC derived precision signal receivedfrom a radio or a satellite system. Currently these requirements can be met with either a cesiumbeam system or a GPS receiver system. Cesium standards are a standalone, highly reliablestratum 1 PRC and provide signal accuracy exceeding the stratum 1 minimum requirement.Cesium is easy to install and turn up when rapid deployment is critical. Minimum training isrequired because it is self contained and simple to operate and maintain. GPS is a reliable PRCbut is dependent on satellite visibility to provide stratum 1 accuracy. The GPS technology doesprovide the service provider time of day using the Network Time Protocol (NTP) since the GPSsystem is traceable to UTC.

    The SSU (G.812) serves as the slave clock in timing nodes that are referenced to an upstreammaster clock. The main purposes of the SSU are:

    o Reference selection: typically more than one source feed is made available to the SSU.

    The SSU must be able to select a reference based on priority, validation, SSM orcombinations thereof;

    o Source validation: performance monitoring of the reference signals in order to

    determine the quality of the best available reference;

    o Holdover: in the event the reference inputs fail, the SSU must maintain an output of a

    specified accuracy over a defined timeframe. This feature is a function of the internalclocks within the SSU.

    o Filtering: the output from the SSU must be impervious to any input signal variations. The

    typical parameters which the SSU will filter are jitter and wander accumulation.

    The SSU has tree main components: inputs, clocks and outputs. Usually the SSU isconfigured with at least 2 input cards for redundancy. This provides the SSU system withmultiple reference input sources and can be of great benefit for performance monitoring andmaintenance.The SSU can distribute multiple outputs, different signal types, and provide outputredundancy. Most manufactures of SSU equipment can provide hundreds of outputs fromthe system, where the most common output type are E1 and 2048 MHz.

    SDH network element clock (SEC) G.813

    About Network Time Protocol (NTP):

    NTP is a protocol that devices can use to obtain accurate time of day (TOD). Accuracy istypically within nanoseconds of UTC.

    NTP is transmitted on top of TCP/IP

    Two operational modes: Client/Server and Broadcast

    Most time servers are GPS

    If synchronization is not guaranteed, considerable degradation in network function, and eventotal failure of the network can be the result. To avoid this worst case scenario, all network elementsneed to be synchronized to a central clock. This central clock is generated by a high precision PrimaryReference Clock (PRC) unit conforming to ITU-T G.811 (accuracy of 10-11). This clock signal must be

    4 of 14

  • 8/3/2019 Transmission - Synchronization Design

    5/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    distributed throughout the entire network. A hierarchical structure is used for this; the signal is passed onby the subordinate Synchronization Supply Units (SSU) and Synchronous Equipment Clock (SEC).

    PRC

    G.811

    SSU

    G.812

    SSU

    G.812

    SEC

    G.813

    SEC

    G.813

    SEC

    G.813

    SEC

    G.813

    Figure 1: Synchronization hierarchy

    The clock signal is regenerated in the SSU and SEC with the aid of phase-locked loops. If the clocksupply fails, the affected network element switches over to a clock source with the same or lower quality,or if this is not possible, it switches to hold-over mode.

    3. Rules

    The SDH frame synchronization method uses a technique called pointer adjustments. The pointer is anaddress field that tells the multiplexer where the VC is located within the overall frame. Adjusting thepointer address to inform the multiplexer that the VC position has moved can rectify any real delays thatoccur over the network. The pointer addresses are moved in blocks of 3 bytes, so each pointermovement adds 3 unit intervals of jitter onto the system.

    A maximum of 15 NE from the PRC can be daisy-chained until it needs to be filtered to thestability set in G.812. After 60 NE, the clock source must be regenerated to achieve a G.811standard.

    External synchronization equipment, often GPS or Cesium based, is used to achieve this in a network.

    ADM/DXC systems must meet the objectives laid down by G.813.

    However, in practical synchronization network design, the number of network elements in tandemshould be minimized for reliability reasons.

    It is recommended that all node clocks and network element clocks are able to recover timing

    from at least two synchronization distribution trails. The slave clock must reconfigure to recovertiming from an alternative trail if the original trail fails.

    5 of 14

  • 8/3/2019 Transmission - Synchronization Design

    6/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    In the event of a failure of synchronization distribution, all network elements need to recovertiming from the highest hierarchical level clock source available.

    To effect this, both G.812 and G.813 clocks may have to reconfigure and recover timing from one of theiralternate synchronization distribution trails. This will ensure that a SDH network element clock-timednetwork element rarely enters holdover or free-run mode. However, it may have to recover timing from aG.812 clock which is itself in holdover if this is the highest hierarchical level source available to it.

    To provide an example of a reconfiguration, if the first network element from the PRC loses itssynchronization trail from the PRC, it must reconfigure and accept timing from the G.812 slave clock.This is shown in the following figure:

    G.811 clock

    1st NE

    G.812 clock

    2nd NE

    Nth NE

    Timing

    propagation

    G.811 clock

    1st NE

    G.812 clock

    2nd NE

    Nth NE

    Timing

    recovery

    Before

    recovery

    After

    recovery

    Figure 2: Synchronization reconfiguration example

    4. PDH and SDH interworking

    In many scenarios, there is a need to carry a PDH payload across multiple SDH islands. There is noabsolute guarantee that every PDH multiplex chain will accept the output jitter appearing at the SDH/PDHboundary. This is because there is no specified lower limit to the corner frequency of the PDHdemultiplex transfer characteristic.

    5. Mobile Base Stations synchronization requirements

    A GSM BTS requires 5 x 10-8. A 3G NodeB requires 2 x 10-8.

    Transmission Transmission

    BSCBTS Retimer

    GPS or Line Sync

    6 of 14

  • 8/3/2019 Transmission - Synchronization Design

    7/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    6. Benefits of a synchronized network

    The overall performance of a digital network is directly related to the degree of synchronization amongstthe network elements and nodes. Among the many benefits of a properly planned synchronizationnetwork, the following are the most prominent:

    Minimizes the number of frame slips on TDM (Time Division Multiplexed) telecom networks

    Minimizes jitter, suppresses short term instabilities and wander resulting in:

    o Clean external reference to network elements

    o Clean synchronization feeds to downstream networks

    Improves operation between PDH and SDH networks

    Fewer customer outages and complaints

    7. Synchronization planning approach and consideration

    In order to design and construct a network that will provide a sound synchronization infrastructure, thefollowing attributes will be accounted for:

    Satisfactory QoS (quality of service) levels for critical services. For example, controlling slip ratesand limiting pointer adjustments;

    Maximized reliability and redundancy with protection schemes;

    Minimize network maintenance;

    Observe hierarchical top down approach;

    Minimize potential timing loops;

    Design a robust network where redundant and seamless protection is available;

    Minimize human interaction;

    Develop a network management scheme for remote central office management.

    With these goals in mind, there are five basic planning considerations for synchronization which isemphasized:

    Maintain hierarchy

    Use the SSU concept

    Avoid timing loops with careful engineering and planning

    Diversity: PRS and cabling diversity

    Minimize cascading effect of distributed timing

    7 of 14

  • 8/3/2019 Transmission - Synchronization Design

    8/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    Designing a synchronization network is a combination of choosing the right clocks and placing them atlogical points in the network to provide maximum performance and reliability.

    7.1 Maintaining hierarchy

    Since synchronization is the result of achieving a common network clock rate, it is important tounderstand the various ways this can be achieved. The objective of the synchronization network is todistribute timing references that are traceable to a primary reference source (PRS) in order to achieve thecommon network clock rate. It is this concept of traceable timing that requires a logical method ofnetworking timing nodes to ensure all nodes are timed by the proper reference points.

    Hierarchical Master Slave

    In a hierarchical arrangement of clocks, the level, or quality of clocks placed in the network is such thatthe higher level clock always provides distribution of the reference signal to lower level clocks. Further, itis sufficient for a site to receive timing from an upstream site of equal type and subsequently pass it on toanother of equal type or level.Historically, this has been the most commonly used network synchronization technique. In thisconfiguration, a reference or master clock is distributed to all nodes through a tree like topology.

    The master-slave synchronization network can be characterized by the following attributes:

    The network consists of hierarchical levels;

    The highest level contains the master clock of the network;

    The master clock provides distribution of the synchronization reference signal to downstreamslave clocks;

    The slave clocks supply timing to yet lower level clocks in the hierarchy;

    The slave clocks in level 2 will serve as a reference to the level 3 clocks if all reference signalscoming from the master are lost;

    Synchronization is always transferred from a higher level to a lower level to avoid timing loops;

    All slave clocks should receive reference signals from multiple sources via different paths.

    Even with the advent of distributed PRS sites, care must be taken in the inter-site distribution of thesynchronization signal. The synchronization signal of the PRS is distributed to the SSUs directly via aring or chain of SECs. These transport signals must be suitable for the transport of synchronization suchthat, at the end of the trail, it must be possible to recover the timing signal which meets the jitter andwander limits (specified in ITU-G.823 and G.825).

    7.2 Use the SSU concept

    The SSU is the key component for deploying proper synchronization in a network. With the ability tovalidate accuracy of an input signal and distribution to all the NE, an SSU will be the mechanism bywhich each and every NE in a single central site should be timed from. The SSU can be a standalonePRC, an integrated PRC/SSU, or use a derived E1 signal.

    8 of 14

  • 8/3/2019 Transmission - Synchronization Design

    9/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    7.3 Avoid timing loops

    Timing loops occur when a clock is being referenced to itself. A condition like this will make the networktiming drift and become unstable and ultimately will cause slips in the network. A potential timing loopexample is shown in Figure. In the figure, when the synchronization output from the SSU is cut, thenetwork will enter into a timing loop.

    SSU SEC

    SECSEC

    SEC

    Figure 2: Timing loop

    7.4 Diversity

    Maintaining network reliability becomes paramount compared to the relative small cost of implementing aredundant PRC clock system such as shown in the next figure.

    GPS Cesium

    Redundant PRC

    Figure: Redundant PRC clock

    7.5 Minimize cascading

    To limit the accumulation of jitter and wander in a synchronization distribution chain, a maximum numberof SSU (Synchronization Supply Unit) and SEC (Synchronous Equipment Clock) are advised not to beexceeded. The maximum number of SSU in a single chain shall not be more than 10, and thenumber of SEC shall not be more than 20 between SSUs, with the restriction that the total numberof SEC within the chain is limited to 60.

    9 of 14

  • 8/3/2019 Transmission - Synchronization Design

    10/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    This explanation is better visualized in the following diagram:

    SSU

    N SEC

    SSU 1

    SSU

    N SEC

    SSU 2

    SSU

    N SEC

    SSU K

    Maximum recommendednodes:K = 10N = 20, with restriction thatSEC clocks do not exceed60 within a single chain

    Figure 3: SSU and SEC limits

    However, there will be cases when some links may have more noise and thus requires an independent

    PRS source at that location. This discretion will be left up to the designer to determine the severity andimportance of the link, and where additional PRS and/or SSU may need to be added.

    8. Synchronization deployment

    Although synchronization networks are different, there are certain constraints that every operator willagree upon, and that is, a synchronization plan must avoid timing loops, minimizes jitter and wander,performance of the network should be enhanced such that yardsticks like dropped phone calls arereduced, and slip are minimized.

    In the following paragraphs, an approach to formulate a synchronization plan will be introduced. Thisapproach adheres to recommendations based on standards, is by no way a firm straight and narrow

    systematic scheme.

    10 of 14

  • 8/3/2019 Transmission - Synchronization Design

    11/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    In practice, passing synchronization in networks is classified in one of two categories, Inter-Node or Intra-Node. In an inter-node scheme, synchronization is passed between nodes usually by way of a transportdevice, such as via SDH. In an intra-node scheme, synchronization is passed from a device such as anSSU to all pieces of equipment co-located in the same site.8.1 Inter-Node synchronization

    With Inter-Node synchronization, distribution of timing from the PRSs to the other clocks in the networkhas, for the most part, been hierarchical. Todays digital network is characterized by a much moredistributed versus hierarchical synchronization topology, due to the advances in clock designs anddistribution techniques.

    It is worthwhile to review the different approaches to network synchronization, in order to fullycomprehend the effects of synchronization. Network elements at different sites can be brought to a

    common clock rate by employing at least one of the following topologies. The most prevalent are:

    Independent PRS;

    Distributed PRS;

    Chains

    Rings

    8.1.1 Independent PRS

    The independent clock approach is based on the plesiochronous approach to network synchronizationwhere data buffers are used to compensate for the frequency differences in the system clocks. Highlystable atomic clock references are used at each node of the network with a buffer large enough tomaintain a specified accuracy over a pre-determined period. Conceptually, independent clocks constitutethe simplest of all techniques for network synchronization. The chief advantage lies in systemsurvivability, where link degradation, or failure of one node does not affect other nodal clocks.

    PRS 1

    PRS 3

    PRS 2

    NETransmission

    link

    Transmissionlink

    Transmissionlink

    NE

    NENE

    Figure 4: Independent clock synchronization concept

    11 of 14

  • 8/3/2019 Transmission - Synchronization Design

    12/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    8.1.2 Distributed PRS

    A distributed PRS approach actually combines the attributes of the hierarchical and independent clockmethods. This technique involves the deployment of multiple PRS types at the same site in an effort toflatten the distribution hierarchy.

    The distributed PRS approach has several advantages:

    It reduces the complexity of the planning effort, i.e. simpler distribution topology;

    Provides for increased performance and reliability;

    Provides an opportunity for multiple alternate references (timing diversity) and supports an active

    surveillance capability;

    Can be used to establish fail-safe timing to SDH ring architectures, independent of softwarebased message techniques;

    8.1.3 Chain

    Synchronization signals can be transported between network nodes via synchronization chains withclock-transparent paths (2 Mbit/s paths) being preferred. The 2 Mbit/s data signals which are part of theSDH payload are not suitable as reference timing signals.

    The main requirements for inter-node synchronization are:

    Transportation of synchronization signals between node clocks must maintain jitter and wanderperformance requirements. In principle, a maximum number of 20 SECs in a chain between twoSSUs will meet the requirements. In practice however, long chains are discouraged and arelimited to 4 to guarantee performance margins;

    SDH NEs require alternative synchronization trails be established in case of distribution linkfailures. Reconfiguration scenarios can be based on SSM (Sync Status Messaging) schemes,which will be explored shortly.

    Under normal operation the synchronization information is transported from SSU 1 which is assumed tobe synchronized to a PRS, to SSU 2 via a chain of NEs containing an SEC clock.

    Under normal operation the network elements between SSU 1 and SSU 2 perform the following tasks inthe transport of synchronization:

    NE 1: Generate a transport signal whose timing is derived from the upstream SSU which issynchronized to a PRS;

    NE 2.n-1: Recover the clock from the incoming transport signal and use it as reference for theSEC. The timing of the outgoing transport signals is derived from the SEC clock;

    NE n: Recover the clock from the incoming transport signal and generate a reference signal forthe SSU.

    In order to keep a lower quality clock from timing a higher quality clock, the reference output signal issquelched when the quality falls below the SSU level (or in some cases, below the PRC level). For that

    12 of 14

  • 8/3/2019 Transmission - Synchronization Design

    13/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    purpose, certain signal criteria must be supervised and the quality information in the SSM is evaluated, ifSSM processing is supported and enabled in the NE.

    Alternatively if SSM is not used during a failure, the NEs in the chain are programmed to acceptsynchronization from the downstream SSU, reversing the direction of synchronization transport up to thepoint of the fault, as shown in the diagram. One important point to add to this scenario is that thedownstream SSU must have a higher quality clock than the SEC internal clocks, otherwise hierarchyprinciples will be violated.

    8.1.4 Rings

    Synchronization rings are typically made up of two parallel chains between SSUs, as shown in thefollowing figure. The synchronization signal is injected into the ring at one of the NEs of the ring via the

    external reference input and transported through one of the SDH transport chains. At some other NE ofthe ring the reference signal is recovered and passed to the downstream SSU.

    A ring creates additional or different synchronization properties, compared to the chain:

    Using automatic reconfiguration mechanisms, each NE in the ring can be reached from theupstream SSU via two synchronization trails. Therefore it is not necessary to injectsynchronization from the downstream SSU into the ring, even under many failure conditions asshown in the previous diagrams;

    The downstream SSU can be synchronized via either of the chains, i.e. via two alternativesynchronization trails. There are two single points of failure, the gateway NEs to the SSUs.Those can be eliminated by injecting and extracting the synchronization signal at two NEs.

    Measures must be taken to avoid timing loops around the ring under certain failure conditions inconnection with automatic reconfiguration. One method is to not close the ring. As shown in thediagram, the SEC clock with the red highlight does not provide a sync reference to the adjacentSEC, thereby not closing the ring.

    Another method to avoid timing loops is to use SSM which is explored next.

    SSM Synchronous Status Messaging

    SSM evolved as a method to transport the quality of a synchronization signal utilizing a message to

    identify the state of the sync signal in order to avoid timing loops and for auto restoration of the syncnetwork when the original sync signal is lost. Or to describe it simply, usage of SSM will allow an SSU ora NE to select the best available sync signal based on the sync quality messge.

    ETSI and ITU have standardized on a SSM messaging scheme to enable this feature. The SSM is codedin the S1 byte of the STM-N section overhead reflecting the SSM quality message.

    The fundamental strategy behind SSM is when a timing failure occurs, the network elements now havethe ability to take an alternate timing reference signal based upon the quality level message that isavailable at its input and auto correct itself.

    Further the network element will relay this information to its downstream network element.

    13 of 14

  • 8/3/2019 Transmission - Synchronization Design

    14/14

    SerieGuidelines

    Title74111877.doc

    Date20/10/2006

    DepartmentTechnical/Engineering

    ServiceTransmission

    8.2 Intra-Node synchronization

    In an intra-node synchronization distribution, the synchronization reference provide a reference signal toall co-located network elements i.e., same site. The reference signals are generated in a node clockwhich is usually a SASE (StandAlone Synchronization Equipment).

    Although some NEs feature an embedded SSU, the SASE implementation is preferred from a planningand maintenance perspective. The SASE is synchronized to a reference signal coming from a PRS viainter-node synchronization transport. [TRANS-SYNC] It is recommended that intra-node synchronizationdistribution is in the form of a logical star.Properties and functionalities for intra-node synchronization:

    Dedicated signals for synchronization transport to and from the node clock:o

    2048 kHz (G.703)o 2048 kbit/s (G.703)

    o 2048 kbit/s (G.704) supporting SSM

    For robustness against faults in the synchronization transport several transport signals arriving atthe node are selected as potential reference sources;

    Ensure that the reference signal at the inputs of the node clock is sufficient. This can be achievedin several ways:

    o Squelch the reference output at the NE if the quality indicated by the SSM message in

    the incoming traffic signal is below a threshold message. The threshold must not belower than the quality of the node clock (in holdover mode, i.e. SSUT). Additionally othersupervision criteria like Loss of Signal, Loss of Frame etc can be used for squelching;

    o Sync signals at the inputs of the node clock with respect to frequency deviation, phasejumps, can be measured with performance monitoring from the SSU.

    Collect the quality information of the NEs in a management system and select the best signal atthe node clock by management command. This solution requires that the NEs and the nodeclock can be accessed from a common management system;

    Provide an SSM message for the STM-N signals leaving the node. In order to insert the correctvalue it is necessary to have the information about the quality of the reference signal in use at thenode clock;

    If the intra-node synchronization distribution is accomplished via 2048 kHz signals, the only way

    to transport information is squelching the signal dependent of some supervision criteria;

    If 2048 kbit/s signals supporting SSM are used for intra-node synchronization, the qualityinformation can be forwarded using the SSM message. This solution requires that all involvedpieces of equipment have 2048 kbit/s reference inputs and outputs also support SSM.

    If none of the enumerated options are available, a fixed SSM message must be inserted into theSTM-N signals leaving the node. The assignment of the appropriate SSM message is a task forthe synchronization planner.

    The synchronization signal can be transported between network nodes via synchronization chains withclock-transparent paths (2 Mbit/s path) being preferred. The 2 Mbit/s signals which are part of the

    SDH payload are not suitable as reference signals because they are subject to wander caused bypointer adjustments.

    14 of 14