Practical test tips for network managers.
Operational and cost benefits make SONET (Synchronous Optical Network) the technology of choice in networking. Ring structures with complex network elements (NEs) are typical. It pays to run an acceptance test to check for proper functionality of the network before bringing these network structures into operation.
Ring structures are gaining popularity among SONET designers today, both in long-haul and local-loop applications, due to their excellent reliability.
A dual ring is used to transport synchronous transport signals (STS), with the outer ring ('working lines') used for traffic and the inner ring ('protection lines') used for backup in case of faults or a drop in signal quality. This prevents a total network failure while ensuring high transmission quality.
Even though manufacturers perform extensive tests on their network elements before delivery, problems can still occur during installation and commissioning of these networks. To ensure proper interplay of the individual system components and error-free transmission, it's useful to know a few test techniques from digital communications technology, such as error and jitter measurements.
Once installed, the SONET ring is put into operation by the system manufacturer and then lined up and tested by the network operator (acceptance process). The most important measurements at this stage are listed in the table.
These measurements must meet the following requirements:
High-speed acceptance process.
Assurance of proper network operation.
Verification of compliance with standard(s).
Key steps to successfully verifying proper ring operation include:
Verifying payload and DCC transparency.
Checking ADM functionality.
Checking ring synchronization.
Ensuring proper APS operation.
Verifying network management functions.
Correct interplay of the entire network is easiest to verify by measuring the payload transparency. A bit error ratio test (BERT) is performed. The ring is opened at a suitable location and a loop connection configured using an optical cable. Using a SONET analyzer, a BERT is then performed in all of the communications channels, i.e., all 12 STS-1 channels (or 4 STS-3c channels) are checked one after another for bit errors in an OC-12 ring.
The generator in the test set provides a pseudo-random bit sequence (PRBS) for use as a test signal that is then inserted as a bulk signal into an STS-1 or STS-3c frame. The receiver detects and counts the bit errors that occurred in the entire tributary and performs error analysis.
To verify the guaranteed quality of service (QoS) is maintained, a 24-hour long-term analysis (ANSI T1.514) is performed. Detecting sporadic errors requires simultaneous recording of anomalies and defects (in addition to the performance evaluation).
Other commissioning measurements deal with built-in TMN (telecommunication management network) functions, which enable software control of all network elements. The data communication channels (DCCs) in the transport overhead (TOH) are used for this purpose (D1 to D3 or D4 to D12). Proper operation of the management functions requires error-free transmission of the overhead bytes. To check for DCC transparency, a BERT is performed in the appropriate channels as described above.
Using a detailed functional test procedure during the line-up phase of a SONET ring, proper configuration and functionality of all network elements is checked, along with the specific functions. This prevents disruptions during 'live' network operation, which are much more expensive to correct later on. Add-drop multiplexers (ADMs) are the most common network elements when building synchronous ring structures. They are capable of dropping asynchronous signals (DS1, DS3) and low-speed synchronous tributary signals from a high-speed SONET bit stream.
The add/drop functionality can be tested using a BERT, but the tributary ports must be looped back on each ADM.
During normal operation, all of the network elements in a ring work with a uniform central clock. One network element is synchronized to an external central clock frequency, while the other NEs derive their clock from the line clock.
Clock derivation for the individual NEs must be configured via network management according to the priority scheme in use and the redundancy paths must be taken into account. In case of faulty configuration, the ring will not be synchronous. This can result in jitter and/or long-term drift (wander), with the clock quality suffering. Jitter analysis, pointer analysis, and wander analysis are necessary to isolate such problems.
Jitter analysis: An initial insight into the synchronization status is offered by checking the output jitter of a tributary. In case of faulty synchronization, jitter values ranging up to several UI can occur. The measurement is configured as the test signal passes through each ADM. In case of excessive jitter values (indicating faulty configuration of the network elements), it sometimes helps to have a look at the clock derivation priority table.
Pointer analysis: Jitter analysis alone does not provide a complete picture of the clock quality. For example, detecting long-term drift requires finer analysis. Therefore, it is necessary to get both the direction of pointer activity and the distribution vs. time. An analyzer is attached to the ring right before the externally synchronized NE to allow its test signal to pass through each of the NEs. This ensures that all influences of the NEs connected to the network clock are captured as part of the analysis. If the clock quality is high, it is necessary to record a very small number of single pointer movements with excellent timing precision. For a valid result, the measurement should extend over at least 24 hours.
If the analysis turns up a large number of pointer movements or pointer movements with a discontinuous distribution vs. time, the faulty network element must be isolated.
Wander analysis: When accuracy requirements are very high, long-term wander analysis is imperative. The analyzer is attached to the ring via an optical power splitter. Due to the high absolute accuracy, the analyzer should be synchronized to an external cesium reference. The measurement runs 24 hours, with simultaneous performance analysis and recording of anomalies and defects. If wander components are detected, then the source must be determined.
In the case of an interruption in a SONET ring, automatic switching from the faulty working line to the protection line occurs (automatic protection switching, APS). APS is controlled via overhead bytes K1 and K2. It is critical to maintain a switching time of 50 ms. If switching does not comply with this figure, an avalanche effect can result. A wave of alarms occurs and entire link segments or rings can be taken out of operation.
As stimulus for APS, a technique known as block & replace is commonly used. The ring is closed and the SONET tester looped into the working line. In this configuration, the SONET tester receives the optical signal (e.g., OC-48), replaces a synchronous channel, including the TOH, and forwards the modified signal. The receiver 'blocks' (analyzes) the same channel. Depending on the NE configuration, the switching can be triggered by inserting B1, B2, or B3 errors.
If the switching time is violated or switching does not occur at all, the source of the problem can be isolated by analyzing the APS protocol. This involves capturing and analyzing overhead bytes K1 and K2 on the protection line.
Alarm sensors: By checking the alarm sensors, it is possible to determine whether the system components properly recognize error states. Besides detecting the errors, the NEs must also report them to the network management system via the appropriate data communication channels (DCCs). Typically, these functions are only checked on one STS-1 tributary. Defined tributary error states (e.g., LOS) are simulated and proper sensor operation is confirmed (observation of LOS, B2, etc.).
Path trace (routing check): Proper routing of a number of tributaries is possible only if the source can be properly assigned to each payload. This is described in the various hierarchy levels with overhead bytes J1, J2 as a 64-character data string. To check for proper operation, a path trace is generated in a selected DS1 channel and then compared with the expected value. After the path trace is altered, it is checked whether the network element triggers a TIM alarm (trace identifier mismatch, TIM) along with an RDI alarm (remote defect indication, RDI) in the reverse direction.
Successful ring turn-up depends on thorough testing. Combining these five test steps into a well-defined test procedure will ensure success. Futhermore, complete test reproducibility, quality standards adherence, and a standardized, automated test sequence should be used.
Habisreitinger is a product manager for Wandel & Goltermann, Research Triangle Park, N.C.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Technology Information; technologies for testing networks|
|Comment:||Testing methods for synchronous optical networks are presented|
|Date:||Oct 1, 1998|
|Previous Article:||Process sorts good ideas from bad.|