When configuring an MPLS circuit, teams must align technical decisions with business intent to ensure predictable performance and operational simplicity. Asking the right questions up front reduces troubleshooting overhead and prevents costly rework during provisioning or migration.
This guide frames critical considerations around service definitions, topology behavior, and quality assurance so your network team can evaluate options consistently.
| Category | Key Questions | Decision Impact | Verification Artifact |
|---|---|---|---|
| Business Intent | Which applications and sites require strict ordering and latency guarantees? | Drives class of service, QoS, and PE location strategy | Application criticality matrix and performance thresholds |
| Topology & Peering | What CE-PE and Provider-Provider options are available in your footprint? | Determines single homing versus dual homing, load balancing, and resiliency | Topology diagram and peering policy document |
| Service Parameters | Which CoS mappings, bandwidth profiles, and CIR/EIR values are required? | Infijcent shaping, queueing behavior, and SLA enforceability | Service parameters sheet and SLA annex |
| Operations & Governance | What monitoring, alarms, and change control processes will be used? | Drives OAM coverage, troubleshooting SLAs, and responsibility boundaries | Operational runbook and escalation matrix |
Service Level and Class of Service Definition
Clarifying Traffic Expectations and Performance Targets
Define the class of service per application or customer group, including latency, jitter, and loss thresholds, before router policies are set. Coordinate how CoS values map across PE, provider, and CE devices to maintain end to end behavior. Translate business criticality into technical metrics so provisioning tools and operators enforce the right treatment consistently.
Topology, Dual Homing, and Route Control
Selecting PE Locations, CE Options, and Failover Behavior
Choose between single homing and dual homing, active active versus active standby designs, and understand how route targets and label distribution interact with your IGP and BGP. Evaluate whether Option A, Option B, or Option C peering best matches policy isolation and load balancing needs across sites.
Service Parameters, Bandwidth, and SLA Enforcement
Specifying Bandwidth, Burst, and QoS Mappings
Establish CIR, EIR, and PIR values per circuit, map DSCP and MPLS EXP to service classes, and define shaping and policing points to match committed behavior. Confirm queuing and congestion avoidance settings on PE routers so measured performance aligns with what customers are promised in the SLA.
Operations, Monitoring, and Change Management
Defining OAM, Alarms, and Responsibility Boundaries
Determine which MPLS OAM tools, such as FFD or Bidirectional Forwarding Detection, will be used, and align alarm thresholds with service expectations. Document change control steps for provisioning, shutdown, and restoration to minimize risk during maintenance or migration events.
Operational Readiness and Governance
Establish a repeatable checklist that covers service parameters, topology decisions, route control rules, and monitoring coverage before cutover to production traffic.
- Define class of service, bandwidth, and burst parameters for each circuit
- Document PE and CE roles, route targets, labels, and route filtering rules
- Verify topology, dual homing behavior, and failover tests in a controlled window
- Configure MPLS OAM and alarms, and align thresholds with SLA commitments
- Implement change control, rollback steps, and post cutover validation
FAQ
Reader questions
How do I decide on single homing versus dual homing for a new MPLS circuit?
Evaluate the cost of downtime against provisioning and transport costs; dual homing typically adds resiliency but requires careful route target and label strategy to prevent asymmetric paths, whereas single homing simplifies operations at the expense of reduced redundancy.
What is the impact of DSCP to EXP mapping on end to end QoS?
If DSCP values are not consistently mapped to MPLS EXP at the PE and preserved across the provider core, queuing and drop precedence may not match intended service classes, leading to unexpected performance for latency sensitive applications.
How should we handle route targets and VRF labeling in a multi tenant environment?
Use unique route targets per customer VRF to control import and export policies, and coordinate RD uniqueness to prevent route leakage; document label distribution for VPNv4 or VPNv6 prefixes to simplify troubleshooting across provider and customer boundaries.
What metrics should be included in SLA verification for an MPLS circuit?
Include availability by measuring interface and BFD failures, latency and jitter under different load levels, and loss metrics for each service class, with measurement points at CE and at provider demarcation to attribute responsibility clearly.