Label Switched Path (LSP) Data Path Delay Metrics in Generalized MPLS and MPLS Traffic Engineering (MPLS-TE) Networks
Extracted elements (24)
Five distinct metrics (RRFD, RSRD, PRFD, PSFD, PSRD) are defined rather than a single metric to capture the relationship between different signaling events (PATH sent/received, Resv sent/received) and data path readiness, enabling identification of specific implementation behaviors.
RRFD can be a negative value when a node programs cross-connections before forwarding the PATH message downstream, meaning the data path is ready before the Resv reaches the ingress. PRFD is recommended to further characterize this case.
The metrics use Poisson sampling (from the IPPM framework in RFC 2330) to select measurement instances, following the established methodology of related LSP provisioning performance metrics in RFC 5814.
Vendors may program cross-connections in a pipelined manner to reduce overall LSP provisioning delay, causing the data path to not be ready instantly after signaling completes. This gap between signaling completion and data path readiness motivates the metrics defined in this document.
Clock synchronization between ingress and egress nodes is required for RRFD, RSRD, and PSFD measurements, but is NOT required for PRFD and PSRD since both timestamps are taken on the same node.
For rate-based data path connectivity verification (e.g., Ethernet), it is RECOMMENDED that the injecting node use small intervals between successive frames, and the interval value MUST be reported.
If an error-free signal is received by the egress node before a PATH message is sent on the ingress node, an error MUST be reported and the measurement SHOULD terminate.
If the corresponding Resv message is received but no error-free signal is received by the egress node within a threshold period, RRFD MUST be treated as undefined, and the threshold value MUST be reported.
If the corresponding Resv message is sent but no error-free signal is received by the ingress node within a threshold period, RSRD MUST be treated as undefined, and the threshold value MUST be reported.
If the LSP setup fails, metric values (RRFD, RSRD, PRFD, PSFD, PSRD) MUST NOT be counted.
If the PATH message is sent but no error-free signal is received by the ingress node within a threshold, PSRD MUST be treated as undefined, and the threshold value MUST be reported.
In computing the median of a metric, undefined values MUST NOT be included; the median SHOULD be set to undefined if all dT values are undefined or if the sample is empty.
In computing the percentile of a metric, undefined values MUST NOT be included.
PRFD SHOULD be used together with RRFD; a PRFD measurement is RECOMMENDED to be carried out after a negative RRFD value has been observed.
The parameter Ts (maximum waiting time for an error-free signal) MUST be reported along with the LSP forward/reverse data path delay sample.
Failure Ratio is defined as the percentage of undefined metric values (signaling completed successfully but no error-free signal observed) to the total number of requests: Failure Ratio = Number of undefined values / (valid + undefined) * 100%.
PRFD (PATH Received, Forward Data path) measures the delay between when the egress node receives the last bit of the PATH message and when it receives an error-free signal, useful for identifying implementations that program cross-connections before forwarding the PATH message downstream.
PSFD (PATH Sent, Forward Data path) measures the delay from when the ingress node sends the first bit of the PATH message to when the egress node receives an error-free signal on the forward data path, providing an end-to-end view of forward data path setup time.
PSRD (PATH Sent, Reverse Data path) measures the delay from when the ingress node sends the first bit of the PATH message to when the ingress node receives an error-free signal on the reverse data path, complementing PSFD for bidirectional LSP characterization.
RRFD (Resv Received, Forward Data path) is a metric measuring the delay between when the last bit of a Resv message is received by the ingress node and when an error-free signal is received by the egress node, indicating the forward data path is ready.
RSRD (Resv Sent, Reverse Data path) measures the delay between when the egress node sends the last bit of the Resv message and when the ingress node receives an error-free signal on the reverse data path of a bidirectional LSP.
The data path delay sample is parameterized by ID0 (ingress LSR ID), ID1 (egress LSR ID), T0/Tf (time window), Lambda (Poisson rate), Th (LSP holding time), Td (maximum wait for LSP setup), and Ts (maximum wait for error-free signal).
The 'error-free signal' is a data-plane-specific indication of data path connectivity: for packet switching, reception of the first error-free packet; for SDH/SONET cross-connects, disappearance of alarms. Implementations must choose a signal appropriate to the technology being tested.
Measurement parameters must be carefully selected to inject only trivial amounts of additional traffic; excessive traffic can skew results and in extreme cases cause congestion and denial of service. In the data plane, injected traffic MUST NOT exceed the rate specified in the LSP setup request, and wavelength-switched measurements MUST use the specified or negotiated lambda with appropriate power.