Intelligent software adds value to Sonet services
n ROBERT COPITHORNE
Intelligent surveillance software oversees network elements, identifies potential trouble areas and diagnoses problems in fiber-optic Sonet networks, keeping services working and customers content. The software is designed for continuous network monitoring and problem resolution--both increasingly difficult to accomplish via traditional means in today`s complex optoelectronic networks.
Network and service providers looking to differentiate themselves on quality rather than price need to build and maintain networks that enable reliable services. Fiber-optic cabling and Synchronous Optical Network (Sonet) transmission equipment provide the foundation for this differentiation, but intelligent software with surveillance capability is needed for proper execution.
Over the years, telecommunications providers have continuously upgraded their fiber-optic networks and invested heavily in Sonet technology. With open competition legislated by the Telecommunications Act of 1996, future fiber-based services are likely to become commodities. Therefore, network and service providers need to position their offerings as competitive, value-added advantages.
Software-based intelligent surveillance, coupled with Sonet`s built-in performance-monitoring capability, creates service differentiated by quality, reliability and performance rather than price. While any provider can compete for commodity-priced bandwidth, some are expected to offer these higher-margin services for mission-critical operations.
For years, network and service providers clamored for built-in performance monitoring in switches and transmission equipment. Although such monitoring is now available, providers have been unable to capitalize fully on this feature because the data must be analyzed continuously. Without intelligent software to collect, analyze and distill the important information from the huge volumes of data, performance monitoring does not live up to users` expectations.
Without such software, network and service providers responded as expected: They shut down the options or ignored the data output. And while monitoring existed on a per-circuit basis, providers were forced to use external hardware and test equipment to troubleshoot and fix network problems.
Since troubleshooting and testing is an invasive procedure, customer permission is needed to take the defective circuit out of service. For a service class differentiated by quality, these procedures have proven inadequate.
Sonet transmission systems offer some relief. If a portion of the customer`s service rides a Sonet fiber-optic facility, it can be rerouted. But the reroute solution is temporary. It creates additional administrative tracking concerns and increases the urgency to quickly identify and fix the problem in the deactivated services. Moreover, if multiple outages occur in shared Sonet equipment, all served customers could be affected.
The best method is to employ the built-in performance-monitoring capability of Sonet transmission and switching equipment along with intelligent surveillance software. Thus, customer traffic becomes an ongoing test pattern that is monitored continuously for signs of pending trouble.
This shift takes network and service providers out of a reactive response mode, so that customer complaints are no longer the trigger for network testing and action. Since intelligent surveillance involves continuous network monitoring, it relies on distinct data-collection and -analysis software instead of test equipment hardware to anticipate and eliminate potential network trouble.
Looking for trouble
Maintenance studies indicate that two out of every three network outages can be predicted. Equipment in a stage of degradation exhibits characteristic signs that can be identified and isolated. Using predictive analysis rules, intelligent surveillance evaluates circuit performance and provides early warnings of failure conditions, so problems can be anticipated and corrected before outages occur.
The process relies on the intelligent reporting capability of network elements. All newly developed network elements have the ability to monitor and report errors, including instantaneous parameters such as bipolar violations, framing bit errors and cyclic-redundancy-check errors.
An error occurrence of any type creates an errored second, a derived parameter. As a single instance, it might only cause a retransmission, typical in most networks. But, dribbling error patterns can be analyzed and characterized as either spurious or indicative of a pending outage. And because dribbling errors are localized, it`s easier to identify problems before a rash of outage-caused sympathetic alarms are generated throughout the network.
Reducing network outages by two-thirds benefits providers and customers alike. Providers can deliver higher quality and more-reliable services with fewer outages, and customers gain guaranteed and measurable service performance.
Expense reductions can be obtained as well. Avoiding customer service outages eliminates service rebates, and it`s less expensive to dispatch and fix problems before they develop into outages.
For the one-third of outages that are unpredictable--such as those resulting from technician errors or outside interventions such as backhoe cuts--intelligent surveillance software can identify the probable cause, allowing quick and directed dispatches.
Outages generate a multitude of direct and sympathetic equipment alarms. The number of alarms multiplies as Sonet transmission speeds increase, and in today`s new virtual networks, the interrogation process for detecting network troubles can be complex. For service offerings differentiated on quality, however, these traffic and topology complexities cannot be allowed to deter an effective response.
Finding problems quickly requires automated analysis by software, not physical analysis by technicians. Intelligent surveillance software can sift through the multitude of alarms and, using root-cause analysis, identify the outage`s probable cause. This approach reduces dependence on skilled technicians--an expensive resource--and relies on system and service logic to identify trouble locations. In this manner, control centers can dispatch technicians to fix rather than to find problems.
As problem detectors, rules-based inference software engines employ alarm correlation techniques. The engines are programmed with physical and logical system and service structures. The inference engines can sort, correlate, identify, and locate trouble areas. This automated process eliminates time as a critical problem-solving constraint--network operators can respond more quickly and delegate the appropriate repair personnel.
In addition, software-based systems allow customers a real-time network "window" on performance data to assess and verify service guarantees. This continuous report card keeps customers actively involved in their own network management, thereby eliminating time-consuming data gathering and manual reporting. It also improves network quality by identifying chronic problem areas and helps prevent contentious customer/provider situations.
Interexchange carriers already use software-based long-distance calling plans to seize market share. Lacking competition, local exchange carriers have not needed to innovate with software. But the newly competitive communications markets legislated by the Telecommunications Act of 1996 are providing the necessary catalysts. q
Robert Copithorne is founder and chief executive of Clear Communications Corp., Lincolnshire, IL. The company develops intelligent software for predicting and preventing network problems.