Frame ordering mode

port/port.yaml

@@ -53,3 +53,49 @@ components:
           type: boolean
           default: false
           x-field-uid: 2
+        frame_ordering_mode:
+          $ref: '#/components/schemas/Port.Options.FrameOrderingMode'
+          x-field-uid: 3
+    Port.Options.FrameOrderingMode:
+      description: >-
+        Controls how an implementation arranges (interleaves) the frames of multiple
+        flows that share source ports and destination ports.
+
+        A flow transmits frames from its source port(s) to its
+        destination port(s). When multiple flows transmit simultaneously towards a
+        common destination port, the transmit ordering determines how the frames of the
+        different flows are interleaved on the wire from the source ports. This is
+        independent of the per-flow payload sequence used for loss and latency
+        measurement.
+
+      type: object
+      properties:
+        choice:
+          description: >-
+            The type of frame ordering mode.
+          type: string
+          default: no_ordering
+          x-field-uid: 1
+          x-enum:
+            no_ordering:
+              description: >-
+                No specific ordering is applied across flows when frames are
+                transmitted from the source ports. The implementation is free to interleave the
+                frames of concurrent flows towards the destination ports in whatever manner is most
+                efficient. This is the lowest-overhead setting and is appropriate when per-flow
+                transmit sequence integrity is not being measured, for example pure throughput or
+                line-rate saturation, loss-only or latency-only measurements, a single flow from
+                one source port to one destination port, and protocol emulation traffic whose
+                ordering is governed by the protocol itself.
+              x-field-uid: 1
+            rfc2889:
+              description: >-
+                Transmit ordering follows the methodology defined in RFC 2889, "Benchmarking
+                Methodology for LAN Switching Devices". The interleaved frames of flow that share source ports and destination ports are arranged using round-robin
+                sequencing before they are transmitted from the source ports. This prevents a
+                destination port from being overloaded by simultaneous flows arriving from multiple
+                source ports and keeps per-flow sequencing valid under load. It is appropriate for
+                RFC 2889 test types such as fully-meshed (many source ports to many destination
+                ports) forwarding, congestion control / head-of-line blocking (many source ports to
+                one destination port), and maximum forwarding rate measurements.
+              x-field-uid: 2