COLLECTING SAMPLES HIERARCHICALLY IN A DATACENTER

• US 20200136939A1
• Filed: 10/30/2019
• Published: 04/30/2020
• Est. Priority Date: 10/26/2018
• Status: Active Grant

• Current Assignee: Vmware LLC (Broadcom, Inc.)
• Original Assignee: VMware, Inc. (Broadcom, Inc.)
• Inventors: Rastogi, Gaurav, Gupta, Ashutosh
• Granted Patent: US 11,171,849 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 11/3006 where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems multiprogramming arrangements G06F9/46; allocation of resources G06F9/50
• CPC Class Codes: G06F 11/3082 the data filtering being achieved by aggregating or compressing the monitored data
• CPC Class Codes: G06F 11/3419 by assessing time
• CPC Class Codes: G06F 11/3452 Performance evaluation by statistical analysis
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 2201/87 Monitoring of transactions
• CPC Class Codes: H04L 41/22 comprising specially adapted graphical user interfaces [GUI]
• CPC Class Codes: H04L 43/045 for graphical visualisation of monitoring data
• CPC Class Codes: H04L 43/065 related to network devices
• CPC Class Codes: H04L 43/0852 Delays
• CPC Class Codes: H04L 43/0894 Packet rate
• CPC Class Codes: H04L 43/10 Active monitoring, e.g. heartbeat, ping or trace-route
• CPC Class Codes: H04L 47/822 Collecting or measuring resource availability data
• CPC Class Codes: H04L 63/1416 Event detection, e.g. attack signature detection
• CPC Class Codes: H04L 63/1425 Traffic logging, e.g. anomaly detection
• CPC Class Codes: H04L 63/20 for managing network security; network security policies in general filtering policies H04L63/0227
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers

Abstract

Some embodiments of the invention provide a method for collecting metric values relating to operations of a set of one or more resources executing on host computers in a datacenter. In some embodiments, the method hierarchically collects and analyzes samples, with a first set of samples collected and analyzed in the data plane, and a second set of samples collected and analyzed in the control plane by aggregating the samples collected in the data plane. In some embodiments, the data plane includes host computers on which sample collecting engines (e.g., service engines) execute, while the control plane includes a set of one or more servers that obtains sample data collected by the host computers, and aggregates and analyzes this data.

Claims

1. A method of collecting values for a particular operational metric relating to operations of a set of resources executing on host computers in a datacenter, the method comprising:

at a particular host computer,

based on a set of sample values for the particular operational metric, dynamically specifying different locations of bins for a histogram, each bin representing a different range of sample values for the particular operational metric;

providing the dynamically specified bin locations to a plurality of other host computers that also collect values for the particular operational metric;

generating the histogram to specify a number of sample values identified for each of a group of the bins of the histogram; and

providing the histogram to a set of one or more servers to aggregate with histograms from other host computers.

2. The method of claim 1, wherein the particular host computer provides the dynamically specified bin locations to the other host computers because the particular host computer is designated as a master host computer for the particular operational metric.

3. The method of claim 2, wherein a particular service engine executing on the particular host computer performs the specifying, providing and generating, and the particular service engine is designated as a master service engine for the particular operational metric and provides the specified bin locations to other service engines executing on other host computers.

4. The method of claim 3, wherein the service engines are load balancers and the particular service engine is a load balancer that is designated as a master load balancer for an application cluster to which the particular operational metric relates.

5. The method of claim 1, wherein the particular host computer provides the specified bin locations to the other host computers so that all the host computers use a common set of bins to generate their respective histograms.

6. The method of claim 1, wherein the generated histogram is a first histogram, the method further comprising:

at the particular host computer,

based on the first histogram, updating the bin locations;

providing the updated bin locations to the other host computers;

based on the updated bin locations, generating a second histogram that specifies a number of sample values identified for each of a group of the bins of the second histogram; and

providing the second histogram to a set of one or more servers to aggregate with histograms from other host computers.

7. The method of claim 1, wherein each histogram is a histogram with variable-sized bin.

8. The method of claim 1 further comprising receiving updated bin locations from the server set.

9. A non-transitory machine readable medium storing a program for execution by at least one processing unit of a particular host computer, to collect values for a particular operational metric relating to operations of a set of resources executing on host computers in a datacenter, the program comprising sets of instructions for:

based on a set of sample values for the particular operational metric, dynamically specifying different locations of bins for a histogram, each bin representing a different range of sample values for the particular operational metric;

providing the dynamically specified bin locations to a plurality of other host computers that also collect values for the particular operational metric;

generating the histogram to specify a number of sample values identified for each of a group of the bins of the histogram; and

providing the histogram to a set of one or more servers to aggregate with histograms from other host computers.

10. The non-transitory machine readable medium of claim 9, wherein the particular host computer provides the dynamically specified bin locations to the other host computers because the particular host computer is designated as a master host computer for the particular operational metric.

11. A method of collecting values for a particular operational metric relating to operations of a set of resources executing on host computers in a datacenter, the method comprising:

receiving histograms generated at a plurality of host computers based on sample values of collected at the host computers for the particular operational metric;

aggregating the received histograms to generate an aggregate histogram;

storing the aggregate histogram in a data storage that has an interface for accessing data represented by the aggregated histogram in order to assess the operations of the set of resources.

12. The method of claim 11, wherein each of the received histograms is a histogram with variable-sized bins.

13. The method of claim 12, wherein all of the received histograms have a set of bins defined at a common set of locations, and aggregating the received histograms comprises adding numbers stored in each set of corresponding bins in the received histograms.

14. The method of claim 11, wherein the interface is a user interface of a management portal.

15. The method of claim 11, wherein the interface is an application programming interface (API) of the data storage.

16. The method of claim 11 further comprising analyzing the aggregate histogram and sending a communication to at least one host computer based on this analysis.

17. The method of claim 16, wherein the communication is sent to a master host computer that generates a set of locations for a set of bins of the histograms and distributes the generated set of locations to other host computers.

18. The method of claim 17, wherein the communication provides information about the aggregate histogram for the master host computer to use in assessing whether the set of bin locations should be modified and distributed to the other host computers.

19. The method of claim 17, wherein the communication provides directions to the master host computer to modify the set of bin locations.

20. The method of claim 16, wherein the communication is to all the host computers and provides directions to the host computers to modify a set of locations for a set of bins of the histograms.

21. The method of claim 16, wherein the communication comprises configuration data for adjusting resource allocation on at least one host computer to improve the operation of the set of resources as gauged by the particular operational metric.