Case Study: Large-scale Frame Relay Enterprise Network

The subject company provides information services (stock market information and securities transactions) to several hundred broker locations nationally. The router-based network is composed of a number of information distribution centers served by T1 Frame Relay. The majority of broker services customers have 56 Kbps or 64 Kbps Frame Relay circuits, while a few have Frame Relay operating over a fractional-T1 access line. Some broker services customers are served using synchronous modems off a T1 channel bank. Backbone links between information distribution centers are dedicated T1 access lines. Because of the time-critical nature of the primary application for the network, the subject company has invested considerably in redundancy and diversity for LAN/WAN transmission and switching equipment.

Why Visual UpTime™?

The subject company acquired the Visual UpTime™ WAN Service Level Management System to improve its time to respond to Frame Relay service outages and abnormalities. The monitoring tools used by their daily operations staff in most cases identified link failures, but could not identify whether the failure was attributable to the Frame Relay provider or to Frame Relay terminating equipment. Restoration of service following a failure was often hampered by the all-too-familiar finger-pointing between a customer and a service provider.

The subject company also needed detailed performance and event correlation data to pro-actively deal with latency, data loss, and to identify changes in information flows. They were reluctant to assign skilled staff away from daily operations to program their existing systems to collect and process performance data to generate reports for analysis, and unwilling to burden WAN circuits and routers with the considerable polling required to collect performance data.

An Overview of the Subject Enterprise Network

The primary application for this WAN is a UDP-based broadcast of stock exchange "ticker" information from distribution centers to brokerage locations. When designing the topology for their delivery and backbone networks, the subject company studied the individual and composite information flows created by this primary application and determined the outbound traffic flow from this broadcast application to be a fairly constant 16-22 Kbps. Currently, very little traffic flows into the distribution centers from customer sites, so the Frame Relay committed information rates (CIRs) are set to 32 Kbps outbound from distribution center to brokerage location, and 8 Kbps inbound from brokerage location to distribution center. T1 access lines at the distribution centers support between 15 and 45 Frame Relay DLC’s.

Figure 4. Enterprise network topology for Case Study

Daily Operations and the roles for Visual UpTime™

The subject company runs a sophisticated network operations center (NOC). NOC staff use a number of custom and commercial applications for very specific purposes. Daily operations staff rely on a custom application to test IP level connectivity to all brokerage location and backbone routers. Due to the size of the network, several copies of this application must be used to assure that IP connectivity to all the routers in the network is confirmed within a meaningful time span. A commercial SNMP NMS is used for collection of routing information, router performance statistics, and for router configuration.

The same application the subject company provides to brokerage locations also runs on client computers located in the network operations center, allowing the NOC staff to observe the broadcast as it is delivered to customers. A trouble ticketing system, T1 circuit management system (for the private line backbone) and LAN analysis equipment are also used by NOC staff.

The subject company uses Visual UpTime™ to complement these traditional NOC tools. The primary daily applications for Visual UpTime™ in the NOC are real-time event monitoring and troubleshooting.

We first describe how Visual UpTime™ is used today by the subject company, and recommend ways to apply additional UpTime™ features that are not currently used. Use Table 2 to find the title of the section that discusses how service level management issue can be addressed using features of the Visual UpTime™ System:

Service Level Management Function	Visual UpTime™ Feature	See Section Entitled:
Alarm monitoring, trouble isolation, identification	Event Processing Window Troubleshooting Windows	"Visual UpTime™ and Trouble Sectionalization"
Detailed trouble analysis	Troubleshooting Toolset: Access Line Window Access Channel Window	"Troubleshooting Frame Relay using Visual UpTime™ Toolsets"
Real-time traffic analysis	Traffic Capture Window	"Real-time traffic analysis over Frame Relay"
Performance monitoring	Troubleshooting Toolset: PVC Summary Window Protocol Summary Window	"Understanding and Managing traffic flows and behavior"
Long-term traffic and trend analysis	Troubleshooting Toolset: Congestion Activity Monitor Burst Advisor Round Trip Delay Window	"Understanding and Managing FR Congestion" "Managing Bursty Data over Frame Relay" "WAN Delay Analysis"
Capacity and performance planning	Long Term Planning and Reporting toolset	"The Tip of the Iceberg"

Table 2. Navigation Aid for This Report

next...