Message ID
56

Message
Sub-optimal performance detected - network capacity might be significantly impacted

Description
The total capacity measured in the network path is in a range that historically has been proven to be sub-optimal. Typically, minor configuration changes can greatly improve performance.

Recommendations
If the total capacity is deemed to be acceptable, no action is required. If the total capacity is below expectations, and one or more diagnostic messages has been generated, take the recommended course of action for each message. Start with the diagnostic message with the highest certainty value. Additional information and suggested courses of action for different capacity measurements can be found in the table below.

Explanation
Total capacity is the highest transmission rate that you can achieve between sender and receiver when no other traffic is on the line; you can read more about it here. In this particular case, the total capacity measurement corresponds to a value that historically has been proven to be sub-optimal. The table below describes the typical causes of sub-optimal performance for a given range of total capacity.

Possible secondary messages
:
* ‘May be due to use of half-duplex or mixture of half- and full-duplex modes’
* ‘May be due to sub-standard or misconfigured NIC or driver’
* ‘May be due to limiting in the mid-path or a slow-path response at target host’
* ‘May be due to a slow-path response at target host’
* ‘May be due to improper QoS (802.1p QoS) configuration resulting in VLAN tagging mismatch’

Total capacity Typical causes
22 - 43 Mbps
  • If the device under test is a router or switch, the measurement reflects the limitations of testing directly to network devices. Since almost all routers and switches are capable of passing full wire speed, this depressed measurement can be ignored. See [interhop analysis](/Delivery/delivery-diagnostics.html) for more.
  • If the device under test is a computer (e.g., desktop or server), consider upgrading the NIC on both ends of the path. Many 100Mbps PCMCIA NICs and laptop card bus controllers operate in this range.
45 - 48.4 Mbps
  • Might be a poorly performing 100Mbps NIC.
48.5 - 49.9 Mbps
  • The device under test is running at 100Mbps half-duplex. Consider upgrading path to 100Mbps full-duplex.
50 - 65 Mbps
  • If the device under test is a router or switch, the measurement reflects the limitations of testing directly to network devices. Since almost all routers and switches are capable of passing full wire speed, this depressed measurement can be ignored. See [interhop analysis](/Delivery/delivery-diagnostics.html) for more.
  • The path might contain a 100Mbps half-duplex component that is limiting end-to-end performance. This condition can occur when the target is full-duplex, and the monitoring point is half-duplex.
65 - 85 Mbps
  • If utilization is greater than 30%, suspect a down-level driver on the NIC.
  • If utilization is less than 30%, suspect auto-negotiation has been set on the router or switch in the test path, which in turn is limiting end-to-end performance.
  • If total capacity is 67 - 75Mbps, this might be due to improper QoS (802.1p QoS) configuration resulting in VLAN tagging mismatch. QoS settings should be identical on both ends of a link.
160 - 200 Mbps
  • Might be a poorly performing GigE NIC or software teaming on the NIC. Typically GigE NICs perform at 225 - 750Mbps with 1500 byte MTU. With 9000 byte MTU, performance can be as high as 900Mbps.