zhou-debs-2011 (2 pages)

Previewing page 1 of 2 page document View the full content.
View Full Document

zhou-debs-2011



Previewing page 1 of actual document.

View the full content.
View Full Document
View Full Document

9 views

Unformatted text preview:

Poster Towards an Inexact Semantic Complex Event Processing Framework Qunzhi Zhou Yogesh Simmhan Viktor Prasanna Department of Computer Science University of Southern California Ming Hsieh Department of Electrical Engineering University of Southern California Ming Hsieh Department of Electrical Engineering University of Southern California qunzhizh usc edu simmhan usc edu ABSTRACT Complex event processing CEP deals with detecting realtime situations represented as event patterns from among an event cloud The state of the art CEP systems process events as plain data tuples and are limited to detect precisely defined patterns Emerging application areas like optimization in smart power grids require CEP to incorporate semantic knowledge of the domain for easier pattern specification and detect inexact patterns in the presence of uncertainties In this paper we present motivating use cases discuss limitations of existing CEP systems and describe our work towards an Inexact Semantic Complex Event Processing InSCEP framework Categories and Subject Descriptors H 4 Information Systems Applications Miscellaneous General Terms Design Languages Keywords Complex event processing Semantic Web demand response 1 INTRODUCTION Complex event processing deals with detecting real time situations represented as event patterns from among an event cloud In recent years research into CEP has received much attention in the research community motivated by applications in domains like financial services 2 and RFID data management 4 Many research prototypes and several commercial systems such as ruleCore 1 and Esper 2 have been developed Demand response optimization DR in Smart Grid is an emerging application area for CEP 5 Smart Grid is the modernization of power grid by integrating digital and information technologies with the deployment of millions of sensors and smart meters to monitor energy use activities 1 2 http www rulecore com http esper codehaus org Copyright is held by the author owner s DEBS 11 July 11 15 2011 New York New York USA ACM 978 1 4503 0423 8 11 07 prasanna usc edu DR a cornerstone application of Smart Grid deals with curtailing power load when a peak load is encountered Continuous data relevant to DR emanating from various sources can be abstracted as events These may be from smart appliances ThermostatChange event smart meters MeterUpdate event weather phenomena HeatWave event or consumer activity ClassSchedule event CEP can correlate these heterogeneous events to detect patterns that predict peak load occurrences or identify load curtailment opportunities for DR in a timely manner Limitations of existing CEP systems limit their uses in diverse information space like Smart Grid Existing systems process events as relational data tuples As such event patterns can only be defined as a combination of attributes presented in event data In addition most CEP systems only support precise pattern matching without any leeway to relax pattern constraints However uncertainty is an intrinsic feature of real world cyber physical applications where potentially incomplete and even incorrect information exist yet need to be matched within certain bounds An effective CEP solution for DR optimization needs to extend traditional CEP systems in two aspects First it must be extensible to meet the organic growth of the Smart Grid information diversity with the provision to easily model and identify new events and event patterns by both domain experts and non domain users Second it should capture uncertainties of events and relax deterministic event patterns for inexact pattern detection 2 USE CASES We present example DR event patterns for load prediction curtailment and monitoring and use them to illustrate key features that our proposed Inexact Semantic Complex Event Processing InSCEP framework should provide Consider in a campus micro grid the DR application processes information coming from sensors and equipments that report their measurements or operations We have the following patterns i Load Prediction A teaching building consumes 90 of its peak load more than 5 classrooms have high probabilities of increasing from base load according to meter readings class schedules and weather conditions ii Load Curtailment The thermostat in one office room is tuned 5 degrees lower than the average set point of thermostats in the same type of rooms which were tuned in the last 30 minutes iii Load Monitoring Conservative curtailment patterns were applied followed by a sequence of meter readings that indicate power load remains steady or increases Traditional CEP systems define patterns by specifying precise constraints of event data However the above examples illustrate the need to incorporate semantics and flexibility in pattern specification The background knowledge of events from multiple domains e g electrical systems appliances room scheduling etc need to be captured In addition flexibility has to exist to allow a limited number of errors or mismatches to still detect a relevant pattern The need for specify such inexact patterns lies in two reasons 1 component events can be probabilistic due to imprecise or incomplete observations and 2 event pattern itself is uncertain and may have infinite acceptable equivalences For instance in the third example the sequence of meter readings need not strictly remain constant or monotonically increase A small fraction of outsider readings should be tolerated 3 INSCEP SYSTEM DESCRIPTION A formal event algebra is defined for specifying complex events that incorporate semantic and inexact query features Some of the algebraic operators such as Selection Aggregation Projection and Renaming originate from relational algebra and are supported in existing CEP systems 1 3 The added expressive power of our algebra lies in the Semantic and Inexact Selection operators Semantic Selection evaluates pattern constraints based on the semantic equivalence of attribute meanings captured by the event ontology instead of syntactic identical attribute values Inexact Selection selects events and allows a limited number of mismatches to detect relevant patterns A similarity function is associated with Inexact Selection to evaluate relevances between matching patterns and target patterns Inexact and Semantic Pattern Detection Events in InSCEP are represented as RDF triples accompanied with timestamps and probabilities Algebraic expressions of complex events are mapped onto an extended version


Access the best Study Guides, Lecture Notes and Practice Exams

Loading Unlocking...
Login

Join to view zhou-debs-2011 and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view zhou-debs-2011 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?