CONFIGURING A SYSTEM TO COLLECT AND AGGREGATE DATASETS

• US 20160226968A1
• Filed: 04/13/2016
• Published: 08/04/2016
• Est. Priority Date: 03/31/2010
• Status: Active Grant

• Current Assignee: Cloudera Incorporated
• Original Assignee: Cloudera Incorporated
• Inventors: Hsieh, Jonathan Ming-Cyn, Robinson, Henry Noel
• Granted Patent: US 10,187,461 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/2023 Failover techniques
• CPC Class Codes: G06F 11/3476 Data logging G06F11/14, G06F11/2205 take precedence
• CPC Class Codes: G06F 16/254 Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses
• CPC Class Codes: H04L 41/12 Discovery or management of network topologies
• CPC Class Codes: H04L 67/1042 using topology management mechanisms
• CPC Class Codes: H04L 67/1087 using cross-functional networking aspects

Abstract

Methods for configuring a system to collect and aggregate datasets are disclosed. One embodiment includes, identifying a data source in the system from where dataset is to be collected, configuring a machine in the system that generates the dataset to be collected, to send the dataset to the data source, identifying an arrival location where the dataset that is collected is to be aggregated or written, and/or configuring an agent node by specifying a source for the agent node as the data source in the system and specifying a sink for the agent node as the arrival location.

Claims

1. A method for configuring a system to collect and aggregate datasets, wherein the system comprises agent nodes to collect the datasets, collector nodes to receive the datasets from the agent nodes, and master nodes configured to change a topology among the nodes in the system, the method being executed by a master node and comprising:

identifying a data source from which a dataset is to be collected;

configuring a machine that generates the dataset to send the dataset to the data source;

identifying an arrival location where the dataset is to be aggregated;

configuring, based on system changes, an agent node in an agent tier; and

configuring, based on the system changes, a collector node in a collector tier, wherein the agent node and the collector node function as peers in a peer-to-peer network.

2. The method of claim 1, wherein configuring the agent node in the agent tier comprises:

specifying a source for the agent node as the identified data source in the system; and

specifying a sink for the agent node as a collector source of a collector node in a collector tier.

3. The method of claim 1, wherein configuring the collector node in the collector tier comprises:

specifying the collector source of the collector node in the collector tier as the identified arrival location; and

specifying a collector sink of the collector node in the collector tier as a distributed file system in a storage tier.

4. The method of claim 1, wherein the dynamically configuring comprises:

performing failover by changing the sink for the agent node to another collector source of another collector node when original collector fails; and

performing fan-out increase by assigning the collector node to receive datasets from another collector node.

5. The method of claim 1, wherein the system changes comprise occurrence of one or more of:

load imbalance, node failure, machine failure, new hardware, or a new collector node.

6. The method of claim 1, wherein the system includes ra ultiple machines which generate datasets that are to be collected.

7. The method of claim 1, wherein the distributed file system includes the Hadoop distributed file system (HDFS).

8. The method of claim 1, wherein the agent node is configurable to perform analytics on the dataset.

9. The method of claim 1, further comprising launching a web page on a machine to access a master for the system using the master'"'"'s universal resource locator (URL).

10. The method of claim 9, further comprising scaling the system by implementing multiple masters on multiple machines.

11. The method of claim 10, further comprising maintaining consistency among the multiple masters using a consensus protocol for ordered state changes.

12. The method of claim 11, wherein the ordered state changes include data flow configurations and node-to-machine mappings.

13. The method of claim 10, further comprising maintaining consistency among the multiple masters using a gossip protocol.

14. The method of claim 13, wherein the gossip protocol is used to maintain consistency among data including one or more of, node status updates, reporting information, and metrics information.

15. The method of claim 9, wherein configuration of the agent node is performed by accessing a link on the web page for the master.

16. The method of claim 1, wherein the data source includes, one or more of, a syslog port, an incoming network connection, and an IRC channel.

17. The method of claim 1, wherein the data source includes, a static file source.

18. The method of claim 1, wherein the data source includes, one or more of, a status update and a feed.

19. The method of claim 1, wherein the data source includes, output from execution of a program.

20. The method of claim 1, wherein the data sink includes one or more of:

a network connection, a physical disk, or a console.

21. A method for configuring a system having multiple machines to collect datasets from the multiple machines and to perform analytics on the datasets, wherein the system comprises agent nodes to collect the datasets, collector nodes to receive the datasets from the agent nodes, and master nodes configured to change a topology among the nodes in the system, the method being executed by a master node and comprising:

identifying data sources on the multiple machines from where datasets are to be collected;

configuring the multiple machines that generate the datasets to send the datasets to the data sources;

identifying an arrival location where a dataset is to be logged; and

providing an interface for accessing the master node to specify, based on system changes, configurations for the multiple machines, the configurations including sources and sinks for the agent nodes and the collector nodes,

wherein the agent nodes and the collector nodes function as peers in a peer-to-peer network.

22. The method of claim 21, wherein the specifying of configurations comprises:

specifying sources for agent nodes in an agent tier as the identified data sources, wherein each agent node is associated with one of the multiple machines;

specifying a sink for each of the agent nodes in the agent tier as a collector source of a collector node in a collector tier;

23. The method of claim 22, wherein the specifying of configurations further comprises:

specifying the collector source of the collector node in the collector tier as the identified arrival location;

specifying a collector sink of the collector node as a distributed file system in a storage tier.

24. The method of claim 23, wherein at least one of the agent nodes of the agent tier or the collector node of the collector tier performs analytics on the datasets.

25. The method of claim 21, wherein the master is used to configure and dynamically reconfigure the agent nodes.

26. The method of claim 21, wherein statuses of the agent nodes are visible via the master.

27. The method of claim 21, further comprising launching a web page on a machine to access the master using the master'"'"'s universal resource locator (URL);

wherein configuration of the agent nodes for the multiple machines is performed by accessing a link on the web page for the master.

28. The method of claim 21, further comprising configuring the sink to be a fan-out sink that sends incoming datasets to child sinks coupled to the sink for data replication or for splitting off a CPU intensive tap.

29. The method of claim 21, further comprising configuring the sink as a failover sink that specifies alternate collectors in case the collector fails.

30. The method of claim 29, wherein the sink is a failover sink that specifies a local disk to store data until the sink that has failed recovers.

31. A system to collect and aggregate datasets, the system comprising agent nodes to collect the datasets, collector nodes to receive the datasets from the agent nodes, and master nodes configured to change a topology among the nodes in the system, wherein a master node in the system has one or more processors configured to perform a method comprising:

identifying a data source from which a dataset is to be collected;

configuring a machine that generates the dataset to send the dataset to the data source;

identifying an arrival location where the dataset is to be aggregated;

configuring, based on system changes, an agent node in an agent tier; and

configuring, based on the system changes, a collector node in a collector tier, wherein the agent node and the collector node function as peers in a peer-to-peer network.

32. The system of claim 31, wherein configuring the agent node in the agent tier comprises:

specifying a source for the agent node as the identified data source in the system; and

specifying a sink for the agent node as a collector source of a collector node in a collector tier.

33. The system of claim 31, wherein configuring the collector node in the collector tier comprises:

specifying the collector source of the collector node in the collector tier as the identified arrival location; and

specifying a collector sink of the collector node in the collector tier as a distributed file system in a storage tier.

34. The system of claim 31, wherein the dynamically configuring comprises:

performing failover by changing the sink for the agent node to another collector source of another collector node when original collector fails; and

performing fan-out increase by assigning the collector node to receive datasets from another collector node.

35. The system of claim 31, wherein the system changes comprise occurrence of one or more of:

load imbalance, node failure, machine failure, new hardware, or a new collector node.

36. The system of claim 31, wherein the system includes multiple machines which generate datasets that are to be collected.

37. The system of claim 31, wherein the distributed file system includes the Hadoop distributed file system (HDFS).

38. The system of claim 31, wherein the agent node is configurable to perform analytics on the dataset.

39. The system of claim 31, the method further comprising launching a web page on a machine to access a master for the system using the master'"'"'s universal resource locator (URL).

40. The system of claim 39, the method further comprising scaling the system by implementing multiple masters on multiple machines.

41. The system of claim 40, the method further comprising maintaining consistency among the multiple masters using a consensus protocol for ordered state changes.

42. The system of claim 41, wherein the ordered state changes include data flow configurations and node-to-machine mappings.

43. The system of claim 40, the method further comprising maintaining consistency among the multiple masters using a gossip protocol.

44. The system of claim 43, wherein the gossip protocol is used to maintain consistency among data including one or more of, node status updates, reporting information, and metrics information.

45. The system of claim 39, wherein configuration of the agent node is performed by accessing a link on the web page for the master.

46. The system of claim 31, wherein the data source includes, one or more of, a syslog port, an incoming network connection, and an IRC channel.

47. The system of claim 31, wherein the data source includes, a static file source.

48. The system of claim 31, wherein the data source includes, one or more of, a status update and a feed.

49. The system of claim 31, wherein the data source includes, output from execution of a program.

50. The system of claim 31, wherein the data sink includes one or more of:

a network connection, a physical disk, or a console.