Problem Statement

We have discussed technology and service overviews in the previous chapters; this chapter focuses on the necessary management aspects when deploying MPLS-based and next-generation network (NGN) services.

As network operators converge multiple services, such as IP VPN, voice, ATM, Frame Relay, and Ethernet over MPLS, the ability for service providers to monitor the LSP integrity, characterize LSP properties, and isolate MPLS forwarding problems becomes critical to their capability to offer services that require service level agreement (SLA) commitments. Traffic engineering, Any Transport over MPLS (AToM), and MPLS IP-VPN are examples of services in which the ability to provide SLA testing and LSP integrity checking might be mandatory (as determined by the network operator managing these services). MPLS OAM plays a crucial role in this picture. Traditionally, SLA testing and LSP integrity checking have been done using Internet Control Message Protocol (ICMP) ping and applications, such as Cisco Service Assurance Agent (SAA), now referred to as Cisco IOS IP Service Level Agreement. MPLS management concerns summarized by customers include the following:

• With all the flow through provisioning, the most complicated part is to ensure that the configuration has worked.

• Solving MPLS virtual private network (VPN) connectivity problems is a complex task for Cisco Certified Internetwork Experts (CCIE).

• You must determine whether quality of service (QoS) configuration is network specific or service specific and whether it is complicated to manage and troubleshoot.

• Troubleshooting performance degradation in MPLS/IP networks is the most labor-intensive NOC activity.

• You must know how to transition PVCs from your ATM core to IP/MPLS infrastructure and how to do so in a multivendor network.

• I need help getting better use of IP/MPLS coreparticularly, increased use of existing network bandwidth resource.

Generally speaking, network management and OAM comprise a set of procedures used to diagnose failures; respond to failures; and test, measure, and verify SLAs within a given network. Diagnostics and tests are applicable to both data and control planes, whereas SLA measurement is related more to the data planealthough the combination of the two is often the most comprehensive approach. Having a consistent way of managing and collecting information from the network elementsa way that is access agnostic and possesses a common command line (CLI) interfacebecomes critical when managing fault, configuration, accounting, performance, and security aspects.

Cisco MPLS Embedded Management offers a set of tools that work together to provide complete MPLS fault, configuration, accounting, performance, and security capabilities (FCAPS). The MPLS Embedded Management architecture is shown in Figure 12-1; it and its components are the main topics of this chapter.

Figure 12-1. MPLS Management Architecture

Most MPLS vendor implementations are standards compliant, and, hence, provide incentive for large networks to deploy multiple vendors in the network. Therefore, developing and implementing a common approach to managing and collecting information from the network elements becomes critical to manage FCAPS aspects.

We discuss aspects of the FCAPS model that pertain to MPLS OAM and network management. We commence with fault management mechanisms, such as LSP ping, trace, and virtual circuit connectivity verification (VCCV), for fault diagnosis. We introduce other mechanisms, such as bidirectional forwarding detection (BFD), LSR self-test, and OAM message mapping (for interworking of OAM messages and states) as part of the MPLS OAM toolkit. We also provide an overview of VPN provisioning via the use of The Cisco IP Solution Center (ISC) and examine accounting and performance mechanisms offered by Netflow and Cisco IOS IP Service Level Agreement (Cisco IP SLA).

We conclude the FCAPS section with recommended security mechanisms, such as the use of Message Digest 5 (MD-5) with control plane protocols.

Some organizations are implementing Layer 3 VPNs, namely BGP-VPNs. These require a per-VPN management model for network operations, thus resulting in a more discreet approach to VPN management.

Per-VPN management examples are described in this chapter. We examine SNMP Management Information Base modules (MIB) as applicable to MPLS-based networks for statistics gathering and topology discovery.

Finally, we provide an overview of the relevant drafts and their status as of the writing of this book.