In recent years, the concepts of Artificial Intelligence (AI) and Machine Learning (ML)...
3.12.5 Commit Control configuration
Commit Control algorithm can be enabled and configured on the Commit Control group of pages. It can be enabled or disabled using the toggle On/Off buttons at the top of the form.
If NetFlow is used to collect statistics Routers MUST be configured to export statistics every minute (or as often as possible). Some router models have default export intervals for either inactive or active flows of up to 1800 seconds. Big delays cause IRP to react very slowly to increased load and reduce the effectiveness of Commit Control feature.
The most important parameters for Commit Control are, as follows:
- Commit Control probing queue slots (core.commit_control.probing_queue_size)
- Minimal prefix bandwidth in Mbps (core.commit_control.agg_bw_min)
Providers Overall block displays Commit Control configuration for all providers including their precedence. See details in the figure below.
Individual provider parameters can be adjusted as well as detail regarding current Commit Control changes applied by an IRP instance can be accessed by following the provided links.
Figure 3.12.18: Providers Overall
3.12.5.1 Inbound Commit Control #
Inbound bandwidth control is part of the Commit Control feature. You need to enable Commit Control and further to enable Inbound Commit control to make use of this feature.
Commit control for inbound highlights the relevant configuration parameters:
- whether inbound commit control is on or off
- whether review and moderation feature is enabled or disabled
- what is the bandwidth estimation algorithm for inbound and others
Refer also to peer.X.95th.mode, core.commit_control.inbound.enabled,core.commit_control.inbound.moderated.
Besides the above mentioned settings for Inbound Commit Control separate low and high limits are configured to instruct IRP when to start and when to stop improving and also how to estimate the effects of inbound improvements.
Estimating inbound traffic Internet traffic has high variability and adjacent measurements can differ significantly between them. In these conditions a leveling function helps mitigate the risk of generating excessive numbers of Inbound improvements due to higher traffic variability of some networks. This estimation can be fine tuned by the Inbound bandwidth estimation algorithm parameter that tells an IRP instance what value to use as the basis of the calculation. See the possible base values below:
3.12.5.2 Optimization of transiting traffic #
Optimization of transiting traffic is setup as part of Inbound Commit Control.
The following configuration parameters can be adjusted:
- Transiting traffic toggle that enables or disables the feature (global.inbound_transit)
- Transit Improvements Max sets the maximum number of transit improvements (core.improvements.inbound_transit.max)
- Transit ASNs and Transit prefixes specify those segments of the Internet that IRP monitors and optimizes (bgpd.prefixlist.asn, bgpd.prefixlist.prefixes)
- Transit Improvements TTL min and max set the lower and upper bounds in seconds to keep a transit improvement (core.improvements.inbound_transit.ttl.min, core.improvements.inbound_transit.ttl.max)
Note that traffic belonging to a specific prefix that fragments a large transit prefix is collected as belonging only to the specific prefix and excluded from the larger prefix traffic statistics. If the specific prefix is also filtered from IRP configuration as for example is the case of a /26 prefix, then the specific prefix will be excluded from optimization of transiting traffic and will not show up in any of the subsequent decisions or reports.- Transit Top N prefixes sets the number of transit prefixes that are collected each cycle and considered for optimization (collector.flow.export.inbound_transit.topn)
- Match transit at egress enables or disable statistics collection for transit prefixes when packets exist the network (collector.flow.process_transit_in_outbound)