Primary, Secondary Server And The Fail-Over Algorithm - AKCP FreeRADIUS SP+ User Manual

SP+ Radius User Manual

Primary, Secondary Server and the Fail-over algorithm

In the sensorProbe+, users can setup Primary and Secondary Radius Servers for both
Authentication and Accounting.
For Access-Request, Accounting-Start and Accounting-Stop packages, the sensorProbe+ will
try to send these packages to the primary server first. If the primary server has failed then the
unit will try with the secondary server (if the user enables it in the configuration). This procedure
will repeat every time when the sensorProbe+ tries to send these packages.
For Accounting-Interim-Update package, the sensorProbe+ will try to send this package to the
primary server first. If the primary server has failed then the unit will try with the secondary
server (if the user enables it in the configuration).
For the next Accounting-Interim-Update request, the sensorProbe+ will use simplified Back off
algorithm (E(c) = ((2^c) - 1) / 2) and use the number of failures as c to calculate the amount of
time in minutes from the last attempt to wait before the unit will try to send Accounting-Interim-
Update request to the primary server again.
For example, if the primary server has failed and the secondary server succeeds, then the next
request will be sent to the secondary server without trying the primary server. After the amount
of time (calculated by the simplified Back off algorithm), the sensorProbe+ will try to send the
package to the primary server again.
If the primary server accepts the request, then the number of failures will be set to 0, otherwise
the sensorProbe+ will increase the number of failures by 1 and calculate the amount of time to
wait before trying to send a request to the primary server again.
- 16 -