Optimizing evaluation patterns and data acquisition for stream analytics in resource-constrained wireless environments

• US 9,137,282 B2
• Filed: 11/15/2013
• Issued: 09/15/2015
• Est. Priority Date: 06/30/2010
• Status: Active Grant

• Current Assignee: Telcordia Technologies Incorporated (Telefonaktiebolaget LM Ericsson)
• Original Assignee: Telcordia Technologies Incorporated (Telefonaktiebolaget LM Ericsson)
• Inventors: Misra, Archan, Falchuk, Benjamin, Chowdhury, Atanu Roy
• Primary Examiner(s): DAILEY, THOMAS J
• Application Number:
  US14/081,308
• Publication Number:
  US 20140074982A1
• Time in Patent Office: 669 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: H04L 65/60 Network streaming of media packets
• CPC Class Codes: H04N 21/4516 involving client characteristics, e.g. Set-Top-Box type, software version or amount of memory available
• CPC Class Codes: H04W 28/06 Optimizing the usage of the radio link, e.g. header compression, information sizing , discarding information system modifying transmission characteristic according to link quality by modifying frame length H04L1/0007; dynamic adaptation of the packet size for flow control or congestion control H04L47/365
• CPC Class Codes: H04W 52/0219 where the power saving management affects multiple terminals
• CPC Class Codes: H04W 72/12 Wireless traffic scheduling
• CPC Class Codes: Y02D 30/70 in wireless communication networks

Abstract

Mobile wireless devices may receive data streams from multiple remote sensors. The sensors may have limited power supplies and memory capacity. Aspects of the invention use statistical characteristics of the sensor data streams and the cost of acquiring a single element of each stream to determine what sequence the sensors should send their data streams in. The cost of acquiring the data may be modified dynamically, depending on parameters such as block size. Additional factors, such as a sensor'"'"'s buffer capacity, may limit the amount of stream elements that may be cached and affect the sensors'"'"' stream transmit sequence. The evaluation order may be dynamically modified using an event processing engine, to reflect both changing statistics of underlying sensor stream tuples and time-varying acquisition costs associated with individual streams. This helps to increase in the operational lifetime of the sensors and associated monitoring applications.

Claims

1. An apparatus for managing data streams from sensor devices received by a mobile client device, the apparatus comprising:

a memory configured to store received data from the sensor devices; and

a processor configured to;

identify a plurality of data streams each corresponding to a respective sensor device, each of the plurality of data streams being receivable by the mobile client device;

specify an order for which each of the plurality of data streams is to be processed by an event processing module of the client device;

receive at least one of the plurality of data streams in the specified order;

determine at least a partial result of a query predicate using data from the received at least one of the plurality of data streams; and

if the query predicate is partially resolved, then dynamically modify some property of a process by which a remaining selected set of the plurality of data streams and data are retrieved by the event proceeding module.

2. The apparatus of claim 1, wherein the dynamic modification includes dynamically modifying the order for acquiring or processing data from selected ones of the remaining selected set of data streams based upon the partial result of the query predicate.

3. The apparatus of claim 1, wherein the dynamic modification includes dynamically modifying a given property of the remaining selected set of sensors.

4. The apparatus of claim 1, wherein the processor comprises an event processing module including:

an event engine configured to apply event operator graphs to data tuples of the plurality of received data streams; and

a dynamic query optimizer configured to compute an optimal order in which the data tuples from the plurality of sensor streams should be retrieved for evaluation by the event engine.

5. The apparatus of claim 4, wherein the event processing module further includes an application specification that specifies characteristics of an application, including operator logic and latency toleration.

6. The apparatus of claim 4, wherein the event processing module further includes a set of optimization models used to capture and compute parameters for evaluation by the dynamic query optimizer.

7. The apparatus of claim 4, wherein the dynamic query optimizer is configured to instruct at least one of the sensor devices to delay transmission of its data stream.

8. The apparatus of claim 4, wherein the dynamic query optimizer is configured to:

associate each of the plurality of data streams with a corresponding acquisition cost; and

set an acquisition order for controlling transmissions of the plurality of data streams based on the acquisition costs.

9. The apparatus of claim 8, wherein the acquisition cost for a given one of the plurality of data streams is a function of a size and rate of data generated by the sensor device configured to issue the given data stream.

10. The apparatus of claim 1, wherein the processor is further configured to track any changes to the statistical distribution of the values of the chosen ones of the plurality of data streams, and use the statistical distribution to modify the order of stream acquisition and evaluation.

11. The apparatus of claim 10, wherein the processor is further configured to use a determination of pre-specified contextual states to trigger the modification of the order of stream acquisition and evaluation.

12. The apparatus of claim 1, wherein the dynamic modification includes:

dynamically modifying an order in which the remaining selected set of the plurality of data streams are to be retrieved by the event proceeding module; and

,

dynamically modifying a number of data tuples in each of the remaining selected set of the plurality of data streams that are to be retrieved by the event processing module.

13. The apparatus of claim 1, wherein if the query predicate is fully resolved then do not perform the dynamic modification thereby avoiding a need to acquire or process data from any remaining data streams.

14. The apparatus of claim 1, wherein the processor is further configured to:

determine a total weighted cost for each of the sensor devices to proactively transmit its data in predetermined data chunks, where weighting for the total weight cost is associated with a likelihood of individual tuples in a given data chunk being relevant to the query predicate to be processed by the event processing module and a communication cost involved in acquiring a given data chunk;

sort the total weighted costs for the respective data streams; and

,

instruct the sensor devices to transmit their respective data streams according to the sorted total weight costs.

15. A client device for managing data streams from sensor devices, the client device comprising:

a processor comprising an event processing module configured to;

receive a query which is viewed as a set of query predicates, where each query predicate refers to one or more of the data streams from the sensor devices;

determine an order in which each data stream will be acquired using both statistics and acquisition costs of the data streams;

acquire at least one of the data streams;

apply the set of query predicates to data elements of the acquired at least one of the data streams; and

,

if the set of query predicates are partially resolved then dynamically modify some property of a process by which a remaining selected set of the plurality of data streams transmit their data by determining an order in which each data stream of the remaining selected set of data streams will be acquired using both statistics and the acquisition costs of the remaining selected set of data streams and then apply the set of query predicates to data elements of the acquired remaining selected set of data streams; and

,

if the set of query predicates is resolved, then output a derived event order as a result of applying the set of query predicates to the data elements of the acquired data streams.

16. The client device of claim 15, wherein the dynamic modification includes dynamically modifying a number of data tuples in each of the remaining selected set of the plurality of data streams.

17. A client device for managing data streams from sensor devices, the client device comprising:

a processor comprising an event processing module configured to;

obtain an energy-cost function which represents a total transmission energy consumed by the sensor devices if their data tuples are transmitted proactively in chunks of data tuples;

determine the chunk size for each sensor device utilizing the energy-cost function;

determine a total weighted cost for each sensor device to proactively transmit data in the chunk size;

sort the data streams from the sensor devices in ascending order of the total weighted cost for each sensor device;

initiate retrieval of the chunks from the sensor devices starting with the data stream with the lowest total weighted cost from which data is yet to be acquired;

upon retrieval of the data stream with the lowest total weighted cost, use the retrieved chunks to refine a result of a query being processed; and

,

if the query is not resolved as a result of the retrieved chunks, then return to the second determine operation to select another data stream to be processed per the query and perform the sort operation, the initiate operation and the use operation until the query is resolved.

18. The client device of claim 17, wherein if the query is not resolved as a result of the retrieved chunks, then return to the first determine operation and recomputed the chunk size for remaining sensor devices and then proceed to the second determine operation to select another data stream to be processed per the query and perform the sort operation, the initiate operation and the use operation until the query is resolved.