PDF of this tutorial
As networks become more complex in adding service and applications, management systems must also evolve. New applications and services are changing how networks are managed and maintained. Also, the constantly changing dynamics of a network based on on-demand connections present increasingly complex management requirements.
Advanced management tools also reduce the total operational cost for network monitoring and maintenance. Operations such as spectrum monitoring over the outside plant for DSL compatibility, as well as collection total error seconds per connection must be automated; the personnel required to evaluate these functions manually would be prohibitive to the business model.
One new application that requires an advanced level of management is voice services. Voice, as a data application, has the most stringent requirement for dependency and is the most affected application when the network experiences fault. Voice services are dependent on low-delay, consistent bandwidth throughput. Any interruption in network service is immediately noticeable to voice sessions. Videoconferencing sessions also have a low tolerance to network faults. As videoconferencing combines voice and real-time video functions into a single application, interruptions in service are immediately noticeable.
Network management includes network maintenance. Updated software must be downloaded into subscriber DSL equipment. This function is best performed in an automated fashion—either by DSL modems querying the network for new software to download, or by commanded mass download.
Management Tools
Management tools include automated functions as well as alarm correlation tools. Several examples of automated functions are included in Table 2:
| signal-to-noise-ratio (SNR) testing | Spectral compatibility is essential in maintaining robust, dependable, DSL–based services. As density of DSLs increases in binder groups, the background noise increases between lines. Additionally, other services, such as T1s and high-bit rate digital subscriber line (HDSL), as well as outside interferes such as amplitude modulation (AM) radio and other radio frequency interference (RFI), may reduce the SNR on individual DSL lines. Finally, each DSL line may be impacted differently by interferes in the distribution plant. Testing of noise levels throughout the distribution plant on a constant or scheduled basis is essential for fault prevention. Carriers may identify interference trends before service interruption. |
| bit-error rate (BER) testing | BERs or rates of severely erred cells may increase as a result of conditions that affect many of only a few DSL lines. The ability to monitor error conditions over a number of DSL circuits may be used to isolate plant-impairing conditions. |
In addition to automated loop- and circuit-testing tools, the ability to perform alarm correlation for root-cause analysis is essential for maximum efficiency over the management plane, particularly in cases where one point of failure may generate many redundant errors from all lower elements affected.
One example of multiple alarms generated by a single component failure involves the failure of a primary network uplink. Where the affected network element is capable of transmitting alarms though a path outside of the failed uplink, the central management systems will receive loss of connectivity for all lower elements (individual DSL elements). Alarm aggregation allows for cause analysis of a multitude of faults resulting in a single overarching alarm in cases where the alarms are related.
