Load shedding in a data stream management system

• US 8,909,671 B2
• Filed: 12/23/2010
• Issued: 12/09/2014
• Est. Priority Date: 12/23/2010
• Status: Active Grant

• Current Assignee: Telefonaktiebolaget LM Ericsson
• Original Assignee: Telefonaktiebolaget LM Ericsson
• Inventors: Pallares Lopez, Miguel Angel, Lafuente Alvarez, Luis Maria
• Primary Examiner(s): Pham, Hung Q
• Application Number:
  US13/994,954
• Publication Number:
  US 20130268560A1
• Time in Patent Office: 1,447 Days
• Field of Search: None
• US Class Current: 707/773
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 2209/5022 Workload threshold
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46

Abstract

It is presented a method for notifying at least a first condition in an industrial system by means of a monitoring system in order to draw the attention of an operator monitoring the industrial system to the at least one condition. The method comprises displaying (S1) the first condition in a first portion of a display screen of the monitoring system, determining (S2) by means of eye tracking an area where a user focuses on the display screen, the area where the user focuses differing from the first portion, and alerting (S3) the user of the first condition in order to allow the user to be notified of the first condition in the industrial system. A monitoring system is also presented herein.

Claims

1. A data stream processing system, comprising:

a data stream management system (DSMS) comprising;

a data store comprising a first memory arranged to store data items received via an input data stream in a data store; and

a first processor comprising;

a continuous query execution circuit configured to execute a continuous query against the received data items to generate an output data stream comprising one or more data items; and

a load shedding circuit configured to execute a load shedding process for discarding some of the data items as received by the DSMS or that have been partially processed by the DSMS;

a DSMS client comprising a second processor and a second memory, the second processor comprising;

a communication circuit configured to receive and process the one or more data items in the output data stream generated by the DSMS; and

a processing circuit configured to execute an action when triggered by the processing of the one or more data items in the output data stream, the action comprising sending a signal, by the DSMS client, to a recipient different from the DSMS client; and

a feedback loop arranged to convey a feedback signal to the DSMS notifying the execution of the action by the DSMS client;

wherein the DSMS is further configured to determine at least one rule governing the load shedding process on the basis of the received feedback signal and the stored data items.

2. The data stream processing system of claim 1, wherein:

the continuous query execution circuit is further configured to;

execute, for each of a plurality of sets of the one or more data items in the input data stream, a continuous query against the one or more data items in the set to generate a corresponding set of one or more output data items in the output data stream; and

insert into the output data stream, in association with each of the corresponding sets of output data items in the output data stream, an identifier which allows one or more of the data items in the corresponding set of data items in the input data stream to be identified;

wherein the data store is configured to store the sets of one or more data items received via the input data stream and the identifiers;

the DSMS client and the feedback loop are further configured to feed back to the DSMS the identifier associated with at least one data item in the output data stream whose processing by the DSMS client triggered the execution of the action; and

the DSMS further comprises a learning circuit configured to identify stored data items to be used to determine the at least one rule governing the load shedding process on the basis of the identifiers fed back via the feedback loop.

3. The data stream processing system of claim 2, wherein:

the continuous query execution circuit is further configured to;

read from, or associate with, a data item in each set of the one or more data items in the input stream a corresponding time-based reference; and

insert into the output data stream, as the identifier associated with each of the output data items in the output data stream, the time-based reference read from or associated with the data item in the set used to generate the output data item;

the DSMS client and the feedback loop are further configured to feed back to the DSMS the time-based reference from the received output stream associated with a data item in the output data stream whose processing by the DSMS client triggered the execution of the action by the DSMS client; and

the learning circuit is further configured to identify stored data items to be used to determine the at least one rule governing the load shedding process on the basis of time-based references received via the feedback loop.

4. The data stream processing system of claim 1:

wherein the DSMS client and the feedback loop are further configured to feed back to the DSMS a priority value associated with the action executed by the DSMS client; and

wherein the DSMS further comprises a learning circuit configured to adapt or create a rule governing the load shedding process on the basis of the priority value.

5. The data stream processing system of claim 1, wherein the DSMS is configured:

in an operational-and-learning mode, to execute the continuous query against data items received via the input data stream, and to determine the at least one rule governing the load shedding process on the basis of the received feedback signal when the data rate from the input data stream is below a certain limit that provides substantially no degradation of a Quality-of-Service figure that is required for the continuous query being executed by the DSMS; and

in an operational-and-load-shedding mode, to perform the load shedding process when the data rate from the input data stream exceeds the certain limit, the load shedding process being performed in accordance with the at least one rule determined during the operational-and-learning mode.

6. The data stream processing system of claim 5:

wherein the DSMS client and the feedback loop are further configured to feed back to the DSMS an indicator indicative of a version of processing logic used by the DSMS client to process the output data stream; and

wherein the DSMS is further configured to detect a change in the processing logic used by the DSMS client on the basis of the received indicator and begin operating in the operational-and-learning mode upon detecting that the processing logic used by the DSMS client has changed.

7. The data stream processing system of claim 1, wherein the feedback loop is arranged to provide to the DSMS as the feedback signal the signal sent by the DSMS client in execution of the action.

8. A data stream management system (DSMS) comprising:

a data store comprising a memory arranged to store data items received via an input data stream;

a processor comprising;

a continuous query execution circuit configured to execute a continuous query against data items received via the input data stream to generate an output data stream;

a learning circuit configured to receive a feedback signal notifying the execution of an action by a client of the DSMS, and determine, using the received feedback signal and stored data items, at least one rule governing a load shedding process for discarding some of the data items as received by the DSMS or that have been partially processed by the DSMS, wherein the action comprises sending a signal, by the DSMS client, to a recipient different from the DSMS client; and

a load shedding circuit configured to execute the load shedding process in accordance with the at least one determined rule.

9. The data stream management system of claim 8, wherein:

the continuous query execution circuit is configured to;

execute, for each of a plurality of sets of the one or more data items in the input data stream, the continuous query against the one or more data items in the set to generate a corresponding set of one or more output data items in the output data stream; and

insert into the output data stream, in association with each of the corresponding sets of output data items in the output data stream, an identifier which allows one or more of the data items in the corresponding set of data items in the input data stream to be identified;

wherein the data store is configured to store the sets of one or more data items received via the input data stream and the identifiers; and

the learning circuit is configured to;

receive, as the feedback signal, the identifier associated with at least one data item in the output data stream whose processing by the DSMS client triggered the execution of the action; and

identify the stored data items to be used to determine the at least one rule governing the load shedding process on the basis of the identifiers fed back thereto.

10. The data stream management system of claim 9, wherein:

the continuous query execution circuit is configured to insert into the output data stream, as each of the identifiers, a time-based reference read from or associated with the data item in the set of data items in the input stream used to generate the output data item with which the identifier is associated; and

the learning circuit is configured to;

receive, as the feedback signal, the time-based reference associated with the data item in the output data stream whose processing by the DSMS client triggered the execution of the action by the DSMS client; and

identify stored data items to be used to determine the at least one rule governing the load shedding process on the basis of received time-based references.

11. The data stream management system of claim 8, wherein the learning circuit is configured to:

receive a priority value associated with the action executed by the DSMS client; and

adapt or create the rule governing the load shedding process on the basis of the received priority value.

12. The data stream management system of claim 8, wherein the DSMS is configured:

in an operational-and-learning mode, to execute the continuous query against data items received via the input data stream, and to determine the at least one rule governing the load shedding process on the basis of the received feedback signals when the data rate from the input data stream is below a certain limit that provides substantially no degradation of a Quality-of-Service figure that is required for the continuous query being executed by the DSMS; and

in an operational-and-load-shedding mode, to perform the load shedding process when the data rate from the input data stream exceeds the certain limit, the load shedding process being performed in accordance with the at least one rule determined during the operational-and-learning mode.

13. The data stream management system of claim 12, wherein the DSMS is configured to:

receive an indicator indicative of a version of processing logic used by the DSMS client to process the output data stream;

detect a change in the processing logic used by the DSMS client on the basis of the received indicator; and

begin operating in the operational-and-learning mode upon detecting that the processing logic has changed.

14. The data stream management system of claim 8, wherein the learning circuit is configured to receive, as the feedback signal, the signal sent by the DSMS client in execution of the action.

15. A data stream management system client, comprising a processor and a memory, the processor comprising:

a communication circuit configured to receive one or more data items in an output data stream generated by a data stream management system (DSMS);

a processing circuit configured to process the received data items in accordance with stored processing logic and execute an action when triggered by the processing of one or more of the received data items, the action comprising sending a signal, by the DSMS client, to a recipient different from the DSMS client; and

a feedback signal generation circuit configured to generate a feedback signal notifying the execution of an action by the processing circuit;

wherein the communication circuit is configured to transmit the generated feedback signal to the DSMS.

16. The data stream management system client of claim 15, wherein:

the communication circuit is configured to receive identifiers in the received output data stream, each identifier being associated with a corresponding data item in the output data stream; and

the feedback signal generation circuit is configured to generate the feedback signal comprising the identifier associated with at least one data item in the received output data stream.

17. The data stream management system client of claim 16, wherein the identifiers in the received data stream are time-based references.

18. The data stream management system client of claim 15, wherein the feedback signal generation circuit is configured to generate the feedback signal comprising a priority value associated with the action executed by the processing module.

19. The data stream management system client of claim 15, further comprising a logic version notifying circuit configured to generate and transmit to the DSMS an indicator indicative of a version of the processing logic used by the DSMS client to process the output data stream.

20. The data stream management system client of claim 15, wherein the feedback signal generation circuit is configured to generate, as the feedback signal, the signal sent by the DSMS client in execution of the action.

21. A method of processing stream data in a data stream processing system comprising a data stream management system (DSMS) and a DSMS client, the method comprising:

the DSMS storing data items received via an input data stream in a data store;

the DSMS executing a continuous query against the received data items to generate an output data stream comprising one or more data items;

the DSMS client receiving and processing the one or more data items in the output data stream generated by the DSMS, and executing an action when triggered by the processing of the one or more data items in the output data stream, the action comprising sending a signal, by the DSMS client, to a recipient different from the DSMS client;

the DSMS being provided with a feedback signal notifying the execution of the action by the DSMS client;

the DSMS determining, on the basis of the provided feedback signal and stored data items, at least one rule governing a load shedding process for discarding some of the data items as received by the DSMS or that have been partially processed by the DSMS; and

the DSMS executing the load shedding process in accordance with the at least one determined rule.

22. The method of claim 21, further comprising:

the DSMS executing, for each of a plurality of sets of the one or more data items in the input data stream, the continuous query against the one or more data items in the set to generate a corresponding set of one or more output data items in the output data stream;

the DSMS inserting into the output data stream, in association with each of the corresponding sets of output data items in the output data stream, an identifier which allows one or more of the data items in the corresponding set of data items in the input data stream to be identified;

the DSMS storing the sets of one or more data items received via the input data stream and the identifiers;

the DSMS client feeding back to the DSMS the identifier associated with the at least one data item in the output data stream whose processing by the DSMS client triggered the execution of the action; and

the DSMS identifying stored data items to be used to determine the at least one rule governing the load shedding process on the basis of identifiers which have been fed back by the DSMS client.

23. The method of claim 22, wherein:

the DSMS reading from, or associating with, a data item in each set of the one or more data items in the input data stream a corresponding time-based reference;

the DSMS inserting into the output data stream, as the identifier associated with each of the output data items in the output data stream, the time-based reference read from or associated with the data item in the set used to generate the output data item;

the DSMS client feeding back to the DSMS the time-based reference associated with a data item in the output data stream whose processing by the DSMS client triggered the execution of the action by the DSMS client; and

the DSMS identifying stored data items to be used to determine the at least one rule governing the load shedding process on the basis of time-based references fed back by the DSMS client.

24. The method of claim 21, wherein:

the DSMS client feeding back to the DSMS a priority value associated with the action executed by the DSMS client; and

the DSMS adapting or creating a rule governing the load shedding process on the basis of the priority value.

25. The method of claim 21, further comprising the DSMS operating:

in an operational-and-learning mode, to execute the continuous query against data items received via the input data stream, and to determine the at least one rule governing the load shedding process on the basis of the received feedback signals, when the data rate from the input data stream is below a certain limit that provides substantially no degradation of a Quality-of-Service figure that is required for the continuous query being executed by the DSMS; and

in an operational-and-load-shedding mode, to perform the load shedding process when the data rate from the input data stream exceeds the certain limit, the load shedding process being performed in accordance with the at least one rule determined during the operational-and-learning mode.

26. The method of claim 25, further comprising:

the DSMS client feeding back to the DSMS an indicator indicative of a version of processing logic used by the DSMS client to process the output data stream; and

the DSMS detecting a change in the processing logic used by the DSMS client on the basis of the received indicator;

the DSMS beginning to operate in the operational-and-learning mode upon detecting that the processing logic used by the DSMS client has changed.

27. The method of claim 21, wherein the signal sent by the DSMS client in execution of the action is provided to the DSMS as the feedback signal.

28. A method of processing stream data in a data stream management system (DSMS), the method comprising:

receiving data items in an input data stream;

storing the received data items in a data store;

executing a continuous query against the received data items to generate an output data stream;

receiving a feedback signal notifying the execution of an action by a client of the DSMS, the action comprising sending a signal, by the DSMS client, to a recipient different from the DSMS client;

determining, on the basis of the received feedback signal and the stored data items, at least one rule governing a load shedding process for discarding some of the data items as received by the DSMS or that have been partially processed by the DSMS; and

executing the load shedding process in accordance with the at least one determined rule.

29. The method of claim 28, further comprising:

executing the continuous query against sets of one or more data items in the input data stream to generate a corresponding output data item in the output data stream for each of the sets;

inserting an identifier into the output data stream, in association with each of the corresponding output data items in the output data stream, which identifier allows one or more of the data items in the corresponding set of data items in the input data stream to be identified;

storing the identifiers and the sets of one or more data items received via the input data stream;

receiving, as the feedback signal, the identifier associated with at least one data item in the output data stream whose processing by the DSMS client triggered the execution of the action; and

identifying the stored data items to be used in determining the at least one rule governing the load shedding process on the basis of received identifiers.

30. The method of claim 29, further comprising:

reading a time-based reference from, or associating a time-based reference with, a data item in each of the sets of data items in the input data stream;

inserting, as the indicators, the time-based references which have each been read from, or associated with, a data item in the set of data items in the input stream, into the output data stream in association with the corresponding output data items in the output data stream;

receiving, as the feedback signal, the time-based reference associated with at least one data item in the output data stream whose processing by the DSMS client triggered the execution of the action by the DSMS client; and

identifying stored data items to be used to determine the at least one rule governing the load shedding process on the basis of received time-based references.

31. The method of claim 28, further comprising:

receiving a priority value associated with the action executed by the DSMS client; and

adapting or creating the rule governing the load shedding process on the basis of the received priority value.

32. The method of claim 28, further comprising the DSMS operating in:

an operational-and-learning mode to execute the continuous query against data items received via the input data stream, and to determine the at least one rule governing the load shedding process on the basis of received feedback signals, when the data rate from the input data stream is below a certain limit that provides substantially no degradation of a Quality-of-Service figure that is required for the continuous query being executed by the DSMS; and

an operational-and-load-shedding mode to perform the load shedding process when the data rate from the input data stream exceeds the certain limit, the load shedding process being performed in accordance with the at least one rule determined during the operational-and-learning mode.

33. The method of claim 32, further comprising receiving an indicator indicative of the version of the processing logic used by the DSMS client to process the output data stream, and beginning to operate in the operational-and-learning mode upon detecting on the basis of the received indicator that the processing logic used by the DSMS client has changed.

34. The method of claim to 28, further comprising receiving, as the feedback signal, the signal sent by the DSMS client in execution of the action.

35. A method of processing stream data in a data stream management system client, the method comprising:

receiving data items in an output data stream generated by a data stream management system (DSMS);

processing the received data items by a processing module in accordance with stored processing logic and executing an action when triggered by the processing of one of more of the received data items, the action comprising sending a signal, by the DSMS client, to a recipient different from the DSMS client;

generating a feedback signal notifying the execution of an action by the processing module; and

transmitting the generated feedback signal to the DSMS.

36. The method of claim 35, further comprising:

receiving identifiers in the data stream from the DSMS, each identifier being associated with a corresponding data item in the data stream; and

generating the feedback signal so as to comprise at least one of the identifiers associated with the one or more data items in the received data stream whose processing triggered the execution of the action.

37. The method of claim 36, wherein the identifiers in the received data stream are time-based references.

38. The method of claim 35, wherein generating the feedback signal comprises generating the feedback so as to comprise a priority value associated with the action executed by the processing module.

39. The method of claim 35, further comprising generating and transmitting to the DSMS an indicator indicative of the version of the processing logic used by the processing module to process the output data stream.

40. The method of claim 35, wherein the feedback signal comprises the signal generated in execution of the action.

41. A computer program product stored in a non-transitory computer readable medium for controlling a data stream management system (DSMS), the computer program product comprising computer program instructions which, when executed by a processor in the DSMS, cause the processor to:

receive data items in an input data stream;

store the received data items in a data store;

execute a continuous query against the received data items to generate an output data stream;

receive a feedback signal notifying the execution of an action by a client of the DSMS, the action comprising sending a signal, by the DSMS client, to a recipient different from the DSMS client;

determine, on the basis of the received feedback signal and the stored data items, at least one rule governing a load shedding process for discarding some of the data items as received by the DSMS or that have been partially processed by the DSMS; and

execute the load shedding process in accordance with the at least one determined rule.

42. A computer program product stored in a non-transitory computer readable medium for controlling a data stream management system client, the computer program product comprising computer program instructions which, when executed by a processor in the data stream management system client, cause the processor to:

receive data items in an output data stream generated by a data stream management system (DSMS);

process the received data items by a processing module in accordance with stored processing logic and executing an action when triggered by the processing of one of more of the received data items, the action comprising sending a signal, by the DSMS client, to a recipient different from the DSMS client;

generate a feedback signal notifying the execution of an action by the processing module; and

transmit the generated feedback signal to the DSMS.