Network virtualization operations using a scalable statistics collection framework
View Patent ↗Some embodiments provide a scalable framework to retrieve statistics relating different aggregated entities. In some embodiments, the aggregated entity is a pair of ports. The pair of ports can be physical ports of one or more physical forwarding elements. The pair of ports can be logical ports of one or more logical forwarding elements. In some embodiments, the aggregated entity is a machine or a group of machines. In some embodiments, the aggregated entity is an access control list (ACL).
1. A network comprising:
a set of physical host machines;
a physical machine;
at least one forwarding element operating on a first set of processing units of the set of physical host machines to process packets for the set of physical host machines, wherein each forwarding element stores a set of flow entries used to process packets, wherein each forwarding element collects statistics for the set of flow entries stored in the forwarding element, wherein statistics for a subset of flow entries in the set of flow entries comprise statistics for a particular logical entity, wherein each flow entry in the subset of flow entries comprise at least one tag identifying the particular logical entity, wherein each forwarding element comprises a flow stats exporter to export flow statistics of each tagged flow entry in the set of flow entries; and
a network manager operating on a second set of processing units of the physical machine comprising an aggregator to receive the statistics from each of the flow stats exporters and to aggregate the statistics for the particular logical entity from the statistics for said subsets of flow entries comprising tags identifying the particular logical entity in order to provide total statistics relating to the particular logical entity.
2. The network of claim 1 , wherein each flow stats exporter exports to the aggregator only the flow statistics of each tagged flow, wherein the set of flow entries comprises at least one flow entry that does not comprise a tag.
3. The network of claim 1 , wherein a particular flow stats exporter sends flow statistics to the aggregator periodically in a set time interval.
4. The network of claim 1 , wherein the aggregator aggregates the statistics by adding a difference between statistics in a current report from the flow stats exporters with the statistics in a previously received report from the same flow stats exporter.
5. The network of claim 1 , wherein the aggregator stores the aggregated statistics in a storage, wherein the aggregator is further to retrieve the aggregated statistics from the storage in response to a user request.
6. The network of claim 1 , wherein the aggregator is further to detect discontinuity in the statistics in a current report.
7. The network of claim 6 , wherein the aggregator detects discontinuity by reading a stat counter in the current report and comparing the stat counter with another stat counter from a previous report.
8. The network of claim 6 , wherein the aggregator detects discontinuity by identifying a duration associated with a flow in the current report and determining whether the duration is less than a duration of the flow in a previous report.
9. The network of claim 6 , wherein the aggregator detects discontinuity by calculating a start time of a flow in the current report and comparing the start time with a start time of the same flow from a previous report.
10. The network of claim 6 , wherein the aggregator stores the aggregated statistic along with a last seen statistics to detect discontinuity, wherein the aggregator does not update the aggregated statistics when discontinuity is detected in statistics of the current report but replaces the last seen statistics with the statistic from the current report.
11. The network of claim 1 , wherein the network comprises a plurality of logical elements comprising said particular logical entity, wherein each forwarding element includes a tagging tool to associate a plurality of different flows with different tags that identify each logical entity in the plurality of logical entities.
12. A system comprising:
a set of physical host machines;
a physical machine;
at least one forwarding element operating on a first set of processing units of the set of physical host machines to process packets for the set of physical host machines, wherein each forwarding element stores a set of flow entries used to process packets, wherein each forwarding element collects statistics for the set of flow entries stored in the forwarding element, wherein statistics for a subset of flow entries in the set of flow entries comprise statistics for a particular access control list (ACL), wherein each flow entry in the subset of flow entries comprise at least one tag identifying the particular ACL, wherein each forwarding element comprises a flow stats exporter to export flow statistics of each tagged flow entry in the set of flow entries; and
a system manager operating on a second set of processing units of the physical machine comprising an aggregator to receive the statistics from each of the flow stats exporter and to aggregate the statistics for the particular ACL from the statistics for said subsets of flow entries comprising tags identifying the particular ACL in order to provide total statistics relating to the ACL.
13. The system of claim 12 , wherein each flow stats exporter exports to the aggregator only the flow statistics of each flow that is associated with at least one tag, wherein the set of flow entries comprises at least one flow entry that does not comprise a tag.
14. The system of claim 12 , wherein a particular flow stats exporter sends flow statistics to the aggregator periodically in a set time interval.
15. The system of claim 12 , wherein the aggregator aggregates the statistics by adding a difference between statistics in a current report from the flow stats exporters with the statistics in a previously received report from the same flow stats exporter.
16. The system of claim 12 , wherein the aggregator stores the aggregated statistics in a storage, wherein the aggregator is further to retrieve the aggregated statistics from the storage in response to a user request.
17. The system of claim 12 , wherein the aggregator is further to detect discontinuity in the statistics in a current report.
18. The system of claim 17 , wherein the aggregator detects discontinuity by reading a stat counter in the current report and comparing the stat counter with another stat counter from a previous report.
19. The system of claim 17 , wherein the aggregator detects discontinuity by identifying a duration associated with a flow in the current report and determining whether the duration is less than a duration of the flow in a previous report.
20. The system of claim 17 , wherein the aggregator stores the aggregated statistic along with a last seen statistics to detect discontinuity, wherein the aggregator does not update the aggregated statistics when discontinuity is detected in statistics of the current report but replaces the last seen statistics with the statistic from the current report.
21. The network of claim 1 , wherein the particular logical entity is one of a logical port distributed across a plurality of forwarding elements and an access control list (ACL).
22. The system of claim 12 , wherein the particular ACL represents a set of rules to permit or deny traffic.