The International Organization for Standardization (ISO) Network Management Forum’s FCAPS model is most commonly used to describe the basic tasks of network management. These tasks are defined as Fault isolation, Configuration, Accounting, Performance, and Security. Provider-based network management can perform any of these tasks as either part of a contracted telecommunications service or as an additional contracted service.
Closed and Open Management Systems
A network management system (NMS) may be as scant as merely an ability to manually connect to a remote device for viewing operational statistics, or as complex as a client-server system supporting automated information exchange and conditional external notifications. Network management systems may be closed (proprietary) or open (standardized).
Closed or Proprietary Management Systems
Closed or proprietary management systems are designed to manage a particular device and do not utilize standardized management protocols or management information bases (MIBs) to collect data from the remote device. Proprietary systems are typically developed by the original equipment manufacturer (OEM) and may feature customized graphic real-time visuals of the device.
Open or Standardized Management Systems
Open or standardized management systems adhere to standard management protocols and MIBs (such as the Simple Network Management Protocol or SNMP), and are capable of managing any remote device supporting a standardized agent. The majority of OEMs support SNMP on their devices and also offer enterprise-specific MIBs that represent information particular to their device.
The most common provider model for network management utilizes standardized SNMP and consists of a primary network management station (NMS) managing one or more client devices. In more complex configurations, secondary network management stations may exist as well, either for redundancy or for specific management capabilities. Client devices will host SNMP agent software that will support the exchange of standardized or enterprise-specific MIB information between the NMS and agent.
Two Types of Data: Trap Data and Query MIB Data
There are primarily two ways a remote client device can report data to a network management system. The first is through a query sent to the device from the NMS that requests standardized information in the form of a MIB. The second is through an unsolicited trap from the remote device. Trap data is sent to the NMS when some condition or set of conditions has been met in the remote device. This may include alarm data (overheating), bit error rates thresholds, bandwidth or CPU utilization thresholds, and more. Both types of data may be used to assess current performance, plan future bandwidth requirements, and even to establish a profile of a device that is likely to degrade or fail in the future.
Trusting Your Provider
Most providers prefer to have the ability to directly monitor their equipment on customer premise, often referred to as ‘eyeballs on site.’ This ability includes determining the immediate status of the gear (up or down) via SNMP-based graphic network maps utilizing color to indicate status. Other data available either from queries or traps include operational statistics (raw bytes in/out, bandwidth utilization, uptime, bit error counts) and device alarms indicating a pre-defined condition that needs attention (such as a redundant power supply failure).
Communications with the remote device will typically be ‘in-band.’ That is, communications with the remote device will be over the contracted circuit providing the telecommunications service. Management communications will therefore require customer bandwidth, but usually a very small fraction of the total circuit bandwidth and well worth the byte cost. Having access to this information facilitates good network planning and can prevent minor events from becoming major service-affecting events. A provider builds trust with their customer by taking this proactive approach to network management and making the customer aware of performance and operational issues that need attention.
Monitoring Versus Reporting
Monitoring of remote devices as described above is a real-time function that may or may not include storing the query or trap data. Reporting requires a formatted version of real-time data collected over a defined time period and related to a specific network behavior, such as bandwidth utilization or error counts. Reporting is typically used by the customer for service level verification, bandwidth accounting, network assessment, and planning purposes. It is a valuable tool in both the provider and customer toolkits, and its collaborative use forges solid provider-customer partnerships.
Syringa Networks will include an SNMP-based monitoring function on all Syringa equipment (switches, routers, muxes) on customer premise terminating a Syringa circuit, as part of our proactive approach to network management. We also offer reporting capabilities upon customer request. Our Network Surveillance Center (NSC) is staffed 24-7 and, along with our engineering staff, is an integral part of Syringa network management. For more information, please contact Larry Price at (208) 297-5225.