Performance Measurement for Service Providers: TWAMP in OcNOS

Service providers carry a contractual obligation to deliver specific performance levels — latency, jitter, and packet loss targets defined in SLAs with enterprise and wholesale customers. Meeting these obligations requires more than reactive monitoring: operators need active measurement tools that continuously probe the network and alert when performance degrades, before customers raise a ticket.

TWAMP (Two-Way Active Measurement Protocol, RFC 5357) provides exactly this. It sends test packets between two network nodes and measures round-trip delay, one-way delay (with clock synchronization), jitter, and packet loss — all without dedicated hardware probes.

TWAMP Architecture

TWAMP Configuration in OcNOS

! OcNOS -- TWAMP server (Node B: responds to test sessions)
!
twamp server
  port 862                       ! Standard TWAMP control port
  timer inactivity 900           ! Close idle sessions after 15 min
!
! Optionally restrict to specific senders:
twamp server
  client-list 10.0.0.1           ! Only accept from Node A

! OcNOS -- TWAMP sender (Node A: initiates measurement sessions)
!
twamp sender
  !
  ! Session 1: measure to PE2 (enterprise SLA monitoring)
  session PE2-SLA
    destination 10.0.0.2         ! Node B loopback
    port 862
    dscp 46                      ! EF: measure on high-priority path
    interval 1000                ! Send test packet every 1 second
    packet-size 256              ! Simulate real application traffic
    timeout 5000                 ! 5-second session timeout
  !
  ! Session 2: measure to remote site (bulk data path)
  session REMOTE-BULK
    destination 10.0.0.5
    port 862
    dscp 0                       ! BE: measure best-effort path
    interval 5000                ! Every 5 seconds
    packet-size 1400             ! Large packet for bulk path testing
  !
!
! Start measurement sessions:
twamp sender session PE2-SLA start
twamp sender session REMOTE-BULK start
!
! View results:
show twamp session PE2-SLA
! Output includes: RTT min/avg/max, jitter, packet loss %
!
show twamp session summary
! Overview of all active sessions with current metrics

Integrating TWAMP with SLA Alerting

! OcNOS -- IP SLA tracking: trigger action when delay exceeds threshold
!
ip sla 10
  icmp-echo 10.0.0.2 source-interface lo
  threshold 50            ! Alert if RTT exceeds 50ms
  frequency 30            ! Probe every 30 seconds
!
ip sla schedule 10 start-time now life forever
!
! Track SLA and trigger route change on threshold violation:
track 10 ip sla 10 reachability
  delay down 5 up 10      ! Hold-down timers to prevent flapping
!
! Use track object in routing:
ip route 0.0.0.0/0 10.0.0.2 track 10
! Primary default route only active when SLA target is reachable

• Performance Measurement: Y.1731 and Y.1564
• Flex-Algo with Link Delay Metric
• OcNOS-SP Product Page
• IP Infusion へお問い合わせ

IP Infusion Engineering Team