This topic describes how the Ethernet SLA feature supports measurement of one-way and two-way delay and jitter statistics and one-way FLR statistics.
The Ethernet SLA is a network performance measurement feature that
-
supports measurement of one-way and two-way delay and jitter statistics
-
calculates one-way frame loss ratio statistics, and
-
evaluates network performance by sending packets and storing data metrics.
| Feature Name |
Release Information |
Feature Description |
|---|---|---|
| Enhancement to Ethernet SLA Statistics Measurement |
Release 7.7.1 |
You can now configure the size of bins for the delay and jitter measurement in Ethernet SLA statistics with a width value ranging from 1 to 10000000 microseconds. This enhancement provides granularity to store more accurate results of SLA statistics in the aggregate bins. In earlier releases, you could only configure the width value for the delay and jitter measurement in milliseconds. This feature introduces the usec keyword in the aggregate command. |
The Ethernet SLA feature tracks various metrics to assess network performance:
-
Round-trip delay time—The time for a packet to travel from source to destination and back to source again.
-
Round-trip jitter—The variance in round-trip delay time (latency).
-
One-way delay and jitter—The router also supports measurement of one-way delay or jitter from source to destination, or from destination to source.
-
One-way frame loss—The router also supports measurement of one-way frame loss from source to destination, or from destination to source.
-
Packet loss count records the overall number of measurement packets lost in either direction.
-
Packet corruption events track data integrity issues.
-
Out-of-order events monitor the sequence of received packets.
-
Frame Loss Ratio (FLR) provides the ratio of lost packets to sent packets expressed as a percentage.
Performance metrics
The system maintains counters for packet loss, corruption, and out-of-order packets for each bucket and reports these metrics as a percentage of the total samples. It calculates the minimum, maximum, mean, and standard deviation for delay, jitter, and loss statistics within each bucket. The system also records individual samples or aggregated bins for these statistics. For synthetic loss measurement statistics, the system reports the overall Frame Loss Ratio (FLR) for the bucket alongside individual measurements or aggregated bins. Finally, the system tracks the packet loss count as the total number of packets lost in both directions, while it measures one-way FLR separately for each direction.
Buckets
The bucket is a time interval that captures collected statistics, records results received during a defined period, and organizes measurements into bins and counters when aggregation is enabled.
The system records all results received during a specific time period in the corresponding bucket. When the system enables aggregation, it assigns a unique set of bins and counters to each bucket. These counters include only the results for measurements initiated during the time period that the bucket represents.
Frame Loss Ratios and availability
The Frame Loss Ratio (FLR) is a performance metric that calculates the ratio of lost packets to sent packets, expresses the result as a percentage value, and measures loss separately in each direction.
The system evaluates availability through the following methods:
-
Availability measures network performance over long periods such as weeks or months.
-
The system calculates the proportion of time during which the network experiences prolonged high loss.
Aggregations
The aggregation is a data management process that organizes samples into range-based bins, reduces memory consumption, and provides efficient statistical storage.
The system manages data storage through two primary methods:
-
When the user enables aggregation using the aggregate command, the system creates bins to store counts of samples within specific value ranges defined by the width keyword.
-
This process stores only a counter for each bin, which minimizes memory usage compared to storing individual results.
-
When the user disables aggregation, the system stores each sample separately to provide more accurate statistical analysis at the cost of higher memory consumption.
One-way delay and jitter measurements
The one-way delay and jitter measurements are network performance configuration processes that - require profile configuration using the type cfm-delay-measurement command, mandate time synchronization between local and remote devices, and utilize specific frequency and time-of-day sources for accuracy.
The administrator follows these requirements to ensure accurate measurement results:
-
The administrator configures the SLA profile using the type cfm-delay-measurement command.
-
The system requires both local and remote devices to maintain time synchronization.
-
The administrator selects frequency sources such as BITS, GPS, SyncE, or PTP.
-
The administrator selects Time-of-Day (ToD) sources such as PTP or DTI, as NTP does not provide sufficient accuracy for these measurements.