VNswitch Router Packet Counters |
Introduction |
This document provides an overview of the VNswitch counters and the effect of packets on counters as packets flow through the router. |
Packet Counter Overview |
The VNswitch contains packet counters that allow you to observe the amount and types
of traffic being processed. The counters keep track of sent and received traffic, in
categories that indicate how many packets have reached various outcomes (terminated,
dropped, bridged, routed, flooded, fragmented, and so on). Packet counters exist at four internal layers to help you trace packets as they flow within the VNswitch:
When a packet is received by an entity within a layer, the packet is either dropped, processed, or passed on to one or more entities within the next layer. Dropped packets cause an error counter within that layer to increment. In some situations, such as bridge ports and interfaces, entities are tightly coupled and dropped packets can increment error counters in two layers. The effects of the packets are not seen in counters at any layers it does not reach. Packets that are successfully processed at a layer increment non-error counters within that layer. Packets sent to the next layer increment non-error counters in that subsequent layer as well. Figure 1 illustrates the four layers and shows the relationships between logical interfaces, bridge ports, VIs, and the IP router. The figure shows which counters are incremented for a typical path a packet can take within a VNswitch. |
Packets arriving at the VNswitch enter the physical interface and can then travel
through each of the four layers. Physical interfaces are the connection jacks for cables,
and have a one-to-one or one-to-many (in the case of an ATM physical interface)
relationship with interfaces. Logical interfaces, shown as circles in Figure 1, are the lowest layer where counters are used.
All packets received and sent are counted by logical interface counters. Error counters
for interfaces can catch some basic types of errors appropriate to the level of decoding
the packet has undergone at this point (for example, a bad FCS) or other errors that are
not necessarily associated with a specific higher-level protocol (for example, buffer
overflow). If such an error is detected on an interface, the packet being sent or received
is discarded, and the appropriate interface error counter is incremented. Otherwise, it is
passed to a bridge port (where bridging runs on all interfaces). Packets arriving at a bridge port (dark triangle in Figure 1) are first subject to the effects of bridging. They may be dropped for numerous reasons (destination address filtering, STP port state, and so on), each causing a single bridge error or a dropped packet counter to increment for that port. If a packet is not dropped, its destination address determines whether it is unicast to another port, flooded out all ports, terminated, and/or delivered to routing. If the packet is bridged out other ports, the bridge attempts to translate and enqueue the packet for sending, if necessary. A failure in this process causes a packet to be dropped and the error counter to be incremented for the received port. A success means that the packet is sent out other ports and counted by them as well. If a received packet is not dropped or sent out by bridging, it is terminated (such as an STP BPDU) and/or submitted to routing. VLAN interfaces (VIs) receive all packets destined to routing. VIs are paired one-to-one with VLANs, which are groups of bridge ports. VIs submit packets for routing on behalf of any ports within their VLAN. VI receive counters keep track of the total number of packets submitted to routing from their VLAN. Outbound packets sent by routing also go through a VI for transmission on a VLAN. VI transmit counters increment once for each packet sent by routing, although multiple packets may be sent on one or more ports (whose counters are incremented as well). Packets sent or received on VIs cannot be dropped by the VI. All errors, overflows, and so on, are detected and counted in other layers. Packets reaching the router may be terminated and are counted by routing. The VNswitch IP counters count transmitted, received, and error packets across all VIs and do not display this information on a per-VI basis.
|
Supported Counters |
This section provides a brief description of interface counters, bridge port counters and the relationship between both counter types. For a complete definition of interface counters and bridge port counters, refer to the DIGITAL VNswitch 900 Series Switch Management guide. Interface CountersThe following table describes each interface counter (IC) associated with the VNswitch. Interface counters IC1A, IC2A, IC3A, and IC4A are identical to IC1, IC2, IC3, and IC4, except the A counters represent VI counters. |
Interface Counter Number | Interface Counter Name |
IC1, IC1A | Unicast packets received |
IC2, IC2A | Multicast packets received |
IC3, IC3A | Unicast packets transmitted |
IC4, IC4A | Multicast packets transmitted |
IC5 | Input overflow drops |
IC6 | Input error drops |
IC7 | Input unknown protocol drops |
IC8 | Input congestion control drops |
IC9 | Output overflow drops |
IC10 | Output error drops |
Bridge Port CountersThe following table describes each VNswitch bridge port counter (BC): |
Counter Number | Counter Name |
BC1 | Port restarts |
BC2 | Total frames received by interface |
BC3 | IP frames fragmented |
BC4 | IP frames not fragmented |
BC5 | Frames submitted to bridging |
BC6 | Frames submitted to routing |
BC7 | Frames with unknown destination address |
BC8 | Frames causing learning transactions |
BC9 | Source address filter drops |
BC10 | Destination address filter drops |
BC11 | Protocol filter drops |
BC12 | Address rate limiting drops |
BC13 | Protocol rate limiting drops |
BC14 | Input buffer overflow drops |
BC15 | Input queue overflow drops |
BC16 | Source or destination port blocked drops |
BC17 | Terminating queue overflows |
BC18 | Fragmentation queue overflows |
BC19 | Translate flood queue overflows |
BC20 | Translation failures |
BC21 | Frames sent by bridging |
BC22 | Transmit queue overflows |
BC23 | Transmit errors |
BC24 | Too big to send on port drops |
Counter RelationshipsSome simple relationships exist between interface counters, bridge port counters, and VI counters. For a given packet in Figure 1, the following relationships exist. Refer to the Interface Counters and Bridge Port Counters sections for a complete list of the counters. Receive RelationshipsExample 1The sum of unicast and multicast packets received on an interface is the total of all received packets. It is expressed as: IC1 + IC2 = BC2 Example 2Some packets received on an interface may be dropped before being submitted to bridging. It is expressed as: BC2 ³ BC 5 Example 3Some packets submitted to bridging will be submitted to routing. The rest are either dropped or bridged. It is expressed as: BC5 ³ BC6 Example 4The packets submitted to routing by a ports VI may represent only a portion of the packets received by that VI since the VIs VLAN may contain other ports. It is expressed as: BC6 £ IC1A + IC2A
|
Transmit Relationships |
Example 1Some packets sent by bridging may be dropped at the interface layer. It is expressed as: IC3 + IC4 £ BC20 |