| Description | The statistics capture database.
The statistics capture table contains summarized
information of the results for a conceptual RTT control
row. A rolling accumulated history of this information
is maintained in a series of hourly 'group(s)'. Each
'group' contains a series of 'path(s)', each 'path'
contains a series of 'hop(s)', each 'hop' contains a
series of 'statistics distribution bucket(s)'.
Each conceptual statistics row has a current hourly
group, into which RTT results are accumulated. At the
end of each hour a new hourly group is created which
then becomes current. The counters and accumulators in
the new group are initialized to zero. The previous
group(s) is kept in the table until the table contains
rttMonStatisticsAdminNumHourGroups groups for the
conceptual statistics row; at this point, the oldest
group is discarded and is replaced by the newly created
one. The hourly group is uniquely identified by the
rttMonStatsCaptureStartTimeIndex object.
If the activity for a conceptual RTT control row ceases
because the rttMonCtrlOperState object transitions to
'inactive', the corresponding current hourly group in
this table is 'frozen', and a new hourly group is
created when activity is resumed.
If the activity for a conceptual RTT control row ceases
because the rttMonCtrlOperState object transitions to
'pending' this whole table will be cleared and reset to
its initial state.
When the RttMonRttType is 'pathEcho', the path
exploration RTT requests' statistics will not be
accumulated in this table.
NOTE: When the RttMonRttType is 'pathEcho', a source to
target rttMonStatsCapturePathIndex path will be
created for each rttMonStatsCaptureStartTimeIndex
to hold all errors that occur when a specific path
had not been found or connection has not be setup.
Using this rttMonStatsCaptureTable, a managing
application can retrieve summarized data from accurately
measured periods, which is synchronized across multiple
conceptual RTT control rows. With the new hourly group
creation being performed on a 60 minute period, the
managing station has plenty of time to collect the data,
and need not be concerned with the vagaries of network
delays and lost PDU's when trying to get matching data.
Also, the managing station can spread the data gathering
over a longer period, which removes the need for a flood
of get requests in a short period which otherwise would
occur. |