For more information, see Status Windows{}.
Anytime InterMapper displays traffic for a link, (using the SNMP Traffic Probe, for example) it polls the following variables:
When you set the SNMP Version to SNMPv1, the following variables are queried:
InterMapper examines these two variables to decide whether an interface is up or down:
MIB Variable | OID | SNMP Version |
---|---|---|
ifAdminStatus | 1.3.6.1.2.1.2.2.1.7 | SNMPv1 |
ifOperStatus | 1.3.6.1.2.1.2.2.1.8 | SNMPv1 |
InterMapper examines these variables to detect error conditions:
MIB Variable | OID | SNMP Version |
---|---|---|
ifInDiscards | 1.3.6.1.2.1.2.2.1.13 | SNMPv1 |
ifInErrors | 1.3.6.1.2.1.2.2.1.14 | SNMPv1 |
ifOutDiscards | 1.3.6.1.2.1.2.2.1.19 | SNMPv1 |
ifOutErrors | 1.3.6.1.2.1.2.2.1.20 | SNMPv1 |
When you set the SNMP Version to SNMPv2c, the following variables are queried:
This variable set is used on an initial scan of the device.
This variable set is polled to display statistics for the device's operation.
Note: In the SNMPv2c , the input and output MulticastPkts and BroadcastPkts MIB variables replace NUcastPkts variables of the SNMPv1 probe, which are deprecated. HCOctets replace the regular Octets counters. Pkts and errors still use the MIB-II 32 bit counters.
InterMapper uses ifInOctets and ifOutOctets to compute the Receive and Transmit bytes/second values, respectively.
The Receive and Transmit packets/second numbers are computed using the sum of the (ifInUcastPkts + ifInNUcastPkts) and (ifOutUcastPkts + ifOutNUcastPkts) respectively.
InterMapper queries a device at specified intervals, and requests a number of SNMP MIB variables. To compute utilization, InterMapper does the following:
Q: A customer writes,
"I see the Received Discards/Minute and Percent Err values for an ATM AAL5 interface are non-zero and I would like to know which variables were used, and what calculation was used to arrive at these numbers.
"We are also graphing the Percent Err: value. This figure is showing errors and my Cisco support folks wanted to know which MIB variables go into the calculation of this percentage and how they are combined to create this number."
A: The Percent Err values were computed as follows:
The one-way percent errors under the Receive section are computed by totalling { ifInUcastPkts, ifInNUcastPkts, ifInErrors, ifInDiscards } as follows:
PERCENT ERROR = totalErrors / totalPkts;
where:
totalErrors = dErrs + dDis and
totalPkts = dUcast + dNUcast + totalErrors
and:
dUcast = ifCurrStats.inUcastPkts - ifPrevStats.inUcastPkts
dNUcast = ifCurrStats.inNUcastPkts - ifPrevStats.inNUcastPkts
dErrs = ifCurrStats.inErrors - ifPrevStats.inErrors
dDis = ifCurrStats.inDiscards - ifPrevStats.inDiscards
Note: Either of 'dErrs' or 'dDis' may be forced to zero if you have "IgnoreInterface Errors" or "Ignore Interface Discards" checked.
The one-way percent errors for outgoing
traffic are similarly computed from the
{ ifOutUcastPkts, ifOutNUcastPkts, ifOutErrors, ifOutDiscards }
statistics.
The two-way Percent error number (just below Utilization on the Interface information menu) is the probability given both one-way error percentages that a packet will be lost making the round-trip across the link and back. If the probability of successful transmission is T and the probability of successful receipt is R (and assuming the act of transmission and receive are relatively independent), then the probability of a successful round-trip is T * R. The probability of error is (1 - T*R).
T and R are computed from the complement of the one-way percent errors above.
When running InterMapper on Mac OS X, you need to disable DHCP and PPP and assign a manually assigned static address to the computer running InterMapper.
To disable DHCP and PPP for all interfaces:
The traffic statistics are samples: the numbers do not change until after InterMapper probes the device again.
SNMP only supplies counts of bytes, packets, or errors, etc. that have passed through or occurred in an interface. These counts increment "forever" (or until the counter rolls over to zero like a car's odometer).
During each poll, InterMapper collects the total traffic and computes the difference with the total traffic from the previous poll. It then divides by the amount of time that has passed to compute the rate (per second or per minute).
Technical note: Even when a counter rolls over (e.g., from 999 to 000), InterMapper will compute the traffic rates accurately. Let's say the two successive samples are 995 and 003. InterMapper subtracts the previous count (995) from the new count (003), assumes that the "003" is actually "1003", and gets the proper difference of 8. Although the counters in the SNMP MIB variable are binary numbers, the same arithmetic principles hold. Thus InterMapper can compute these rates accurately.
To compute the elapsed time accurately, InterMapper uses the sysUpTime variable of the device as a timestamp to calculate the time that has elapsed between subsequent two polls. The time elapsed should roughly correspond to the poll interval; however, it is possible for polls to be delayed occasionally so using the change in sysUpTime to measure the elapsed time is more accurate.