S. Chakravarthy, V. Krishnaprasad, E. Anwar, and S. K. Kim. Anatomy of a composite event detector. Technical Report CIS TR-93-039, University of Florida, December 1993.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... have in fact been imple mented recently to provide [his capability an incomplete list of such systems includes [9, 13, 10, 4] These systems feature composite event detection mechanisms, which are based on formalisms such as Finite State Automata [12] Petri Nets [11] or Event Graphs [3]. The problem of formally specifying the semantics of active rules remains largely unsolved. Indeed, giving a formal semantics to active database behavior presents a challenge even when only simple events are involved. For composite event expressions, this problem becomes even more complex, ....

....composite events. However, there are still some limitations. Si multaneous events cannot be handled. Also, it is not clear how the semantics of general negation (as defined in EPL) can be captured by Petri nets. The composite event detection mechanism of Snoop, which is based on Event Graphs, [3] suffers from these limitations as well. Another concept, whose formal definition is missing in previous approaches is that of event histories and the succession of states in the system. This be comes necessary when introducing simultaneous events. Finally, all previous approaches are limited to ....

S. Chakravarthy, V. Krishnaprasad, E. Anwar, and S. K. Kim. Anatomy of a composite event detector. Technical Report CIS TR-93-039, University of Florida, December 1993.

....and composite events, we explain the event detection process, i.e. the way events are recognized and signaled to the module in charge of executing rules. The primitive event detection process is efficient and fine grained as it is placed within the database it is supposed to monitor. As in [3, 4] our approach for detecting composite events is based on an event graph that store only events that make composite events occur. The event graph is the result of merging event trees. A tree represents a composite type expression and may include nodes calculated from the context (of detection) of ....

....3.1 Time and Events Time cannot be dissociated from events [5] The concept of time can be understood in several ways. We will not detail this concept and the way it is incorporated in databases as it is out of the scope of this paper. Readers may find more information on temporal databases in [20, 28, 2, 3]. The solution to represent a point in time is to define a reference moment and a unit (i.e. second, minute, hour, etc. We choose to use a discrete representation: the Gregorian calendar and we consider as in [5] a discrete time domain isomorphic to this calendar. Whatever the time ....

S. Chakravarthy, V. Krishnaprasad, E. Anwar, and S-K. Kim. Anatomy of a Composite Event Detector. Technical Report UF-CIS-TR-93-039, University of Florida, Gainesville, USA, December 1993.

....active databases has an implicit validity restriction because it only considers transaction units [WCL91] As soon as a transaction commits or aborts, the complete set of basic events and situations is forgotten. Some active databases provide explicit means to specify the actuality of events ([CKAK93], Jas94] During runtime numerous incarnations of an event type occur. Basic event incarnations differ in timestamp and context information. Complex situations are derived from a combination of basic events. The decision which event incarnation has to be considered for a potential new complex ....

....is forgotten, and for a next occurrence of this situation only the first incarnation of each constituent event type is considered. Explicit actuality is more flexible. Either the first occurrence is considered ( first ) or the most recent ( last ) or all occurrences ( all ) Sentinel [CM93] [CKAK93] and AIS [Jas94] allow further semantics, which are special cases of all . first and last consider exactly one occurrence, which is a considerable restriction of information. all leads to a cross product of constituent events, and an explosion of potential complex situations. Context ....

Sharma Chakravarthy, V. Krishnaprasad, E. Anwar, and S.-K Kim. Anatomy of a Composite Event Detector. Technical Report TR93-039 University of Florida, 1993.

.... as the issues of temporal databases and active rule languages are now combined and intertwined [21] In previous systems, the operational semantics of composite event detection has been defined using widely different frameworks, such as Finite State Automata [12] Petri Nets [11] or Event Graphs [4]. Eventhough these systems have captured a great deal of the functionality required by active database applications, they are restricted by the limitations of their underlying This work was partially supported by NASA HPCC grant NAG 5 2225. formalisms. Consequently, it is desirable that a general ....

....events. However, there are still some limitations. Simultaneous events cannot be handled. Also, it is not clear how the semantics of general negation (as defined in the next section) can be captured by Petri nets. The composite event detection mechanism of Snoop, which is based on Event Graphs [4] suffers from similar limitations. A simpler approach to overcome the afore mentioned problems is that of using Datalog 1S [3] This is a temporal language that extends Datalog, by allowing every predicate to have at most one temporal parameter (constructed using the unary successor function s) ....

S. Chakravarthy, V. Krishnaprasad, E. Anwar, and S. K. Kim. Anatomy of a composite event detector. Technical Report CIS TR-93-039, University of Florida, December 1993.

....into several cascading triggers. Since this section shows how active data warehouses can be realized, topics related to materializing cubes and managing events in a highly efficient manner are not subject to discussion. Refer to [3, 29, 39] for an in depth discussion on view maintenance and to [7, 18] for efficient algorithms on event management. 6.1 Multidimensional Data Multidimensional data as described by a conceptual dimensional model is represented as a starflake (i.e. combining star and snowflake schemes) in the relational data warehouse as follows: Each dimension level and its ....

S. Chakravarthy, V. Krishnaprasad, E. Anwar, and S.-K. Kim. Anatomy of a Composite Event Detector. Technical Report UF-CIS-TR-93-039, University of Florida, Computer and Information Sciences, December 1993.

....events that are constructed with the operators sequence, disjunction, and conjunction, whereas the latter group comprises composite events constructed with history, negation, and closure. Several di#erent approaches are used for composite event detection including syntax graphs #Deutsch, 1994; Chakravarthy et al. 1993#, Petri nets #Gatziu and Dittrich, 1994#, #nite state automata #Gehani et al. 1992#, and arrays #Eriksson, 1993#. 2.2 Conditions The condition part of a rule is usually a boolean expression, a predicate, or a set of queries, and it is satis#ed if the expression evaluates to true, or all the ....

Chakravarthy S., Krishnaprasad V., Abwar E., and Kim S. K., Anatomy of a Composite Event Detector. Technical Report UF-CIS-TR-93-039, University of Florida, Florida, 1993.

....allegiance relationship, and the condition evaluation transaction can commit or roll back regardless 9 of the parent transaction. The Sentinel project [Lee96] was a follow on to the HiPAC project at the University of Florida. In the former, they developed a rule definition language called Snoop [Chak91, Chak93a, Chak93b]. The Sentinel project introduced several mechanisms for 1) monitoring events in a distributed fashion, 2) communication among applications using event mechanisms, and 3) integrating rules into a database programming language. Another follow on project to HiPAC was Reach [Kud93] developed at the ....

S. Chakravarthy, V. Krishnaprasad, E. Anwar, and S. K. Kim, "Anatomy of a composite event detector", Technical Report UF-CISE-TR-93-39, CISE Department, University of Florida, 1993.

....21 http: www.his.se ida drts techTransfer 1996 deedsInSe.ps.Z 22 http: www.his.se ida drts techTransfer 1996 deedsInSehandout.ps. Z 10 Buchmann et al. 1995] papers covering the REACH project, which is another active real time database project; Chakravarthy and Mishra, 1994, Chakravarthy et al. 1993, Chakravarthy et al. covering the Snoop event detector based on event graphs which is an efficient event monitoring techniqure, and event contexts; Gehani and Jagadish, 1992b, Gehani and Jagadish, 1992a, Jagadish et al. 1992, Gehani et al. 1993] covering the active database ODE ....

....and Garcia Molina, 1992] additional perspectives on the problems of real time databases. ffl Distributed systems. Books: Mullender, 1994] a very good book covering most of the interesting issues in distributed computing. ffl Event monitoring. Articles: Chakravarthy and Mishra, 1994, Chakravarthy et al. 1993, Chakravarthy et al. the snoop event detector; Buchmann, 1994, Buchmann et al. 1995] which gives a good insight in problems concerning monitoring active (real time) databases [Deutsch, 1994] which has a comparison of different event detection algorithms used in active database; ....

Chakravarthy, S., Krishnaprasad, V., Anwar, E., and Kim, S. K. (1993). Anatomy of a composite event detector. Technical Report UF-CIS-TR-93-039, Computer and Information Sciences Dept, University of Florida.

.... of temporal databases and active rule languages are now combined and intertwined [19] Furthermore, the operational semantics of composite event detection in previous systems have been defined using widely different frameworks, such as Finite State Automata [11] Petri Nets [10] or Event Graphs [3]. Eventhough these systems have captured a great deal of the functionality required by active database applications, they are restricted by the limitations of their underlying formalisms. Consequently, it is desirable that a general more abstract execution (operational) formalism is employed, ....

....composite events. However, there are still some limitations. Simultaneous events cannot be handled. Also, it is not clear how the semantics of general negation (as defined in EPL) can be captured by Petri nets. The composite event detection mechanism of Snoop, which is based on Event Graphs [3] suffers from similar limitations. In this paper, we propose to use Datalog 1S , as the basis for defining the formal semantics of complex events in active rules. Datalog 1S [2] is a temporal language that extends Datalog, by allowing every predicate to have at most one temporal parameter ....

S. Chakravarthy, V. Krishnaprasad, E. Anwar, and S. K. Kim. Anatomy of a composite event detector. Technical Report CIS TR-93-039, University of Florida, December 1993.

....of event composition relative to transaction boundaries, as this has not been properly addressed in the literature. 3. 2 Event Composition Relative to Transaction Boundaries Events can be composed using either finite state automata [GJS92] colored) Petri nets [GD93b] or syntax graphs (e.g. [CKA93], Deu94] A crucial issue for the architecture is the composition of events relative to transaction boundaries and the valid execution strategies of rules, depending on the kind of event. This issue has not been properly addressed elsewhere. Events can be: ffl Simple method events (both ....

Chakravarthy, S., Krishnaprasad, V., Anwar, E., Kim, S.-K; Anatomy of a Composite Event Detector. TR-93-039, U. Florida, 1993.

....at the University of Florida, Database Systems Research and Development Center, Gainesville, Florida. It developed directly from results obtained in the HiPAC project. Several papers discuss prototypes on the basis of Zeitgeist and the DARPA Open OODB toolkit. References: CAM93] CHS92] CKAK93] CKAK94] CM91] CM93] CZ93] CHS92] Secondary: AMC93] 2.15 Starburst The Starburst rule extension was developed at IBM Almaden Research Center, San Jose, CA. It is an integral part of Starburst. The Starburst rule system is a production rule system tightly integrated into the ....

....executing rules. References: BOGM92] CW92] 3.7 Architecture Architecture of active databases are addressed in: References: Cha92] CZ93] Jas94] 3. 8 Implementation Issues There are prototypes implemented, and the following papers discuss details of implementations: References: Car92] CKAK93] CN90] Coh89] DM89] GD93b] GJS92a] Han92b] Han92a] HB91] HK87] HK89] KD93] Mor83] RCBB89] WCL91] 3.9 Optimization Optimization of rule execution spans from efficient event detection to efficient data condition evaluation to fast and non redundant rule action ....

S. Chakravarthy, V. Krishnaprasad, E. Anwar, and S.K. Kim. Anatomy of a composite event detector. Technical Report TR93039, University of Florida, Gainsville, Florida 32611, 1993.

....it as a new event. For example, consider the sequence E1 followed by E2, as described in Figure 3. If two instances of event E1, here denoted e1a and e1b, are raised in that order, and then two instances of event E2, called e2a and e2b, are raised, the problem is how to combine those events. In [CKAK93] four different contexts are described. The simplest one, recent, throws away earlier event instances of a given event and uses the most recent. In the example above one composite event should have been raised consisting of e1b together with e2a. The recent context is the context most likely to ....

S. Chakravarthy, V. Krishnaprasad, E. Anwar, and S.-K Kim. Anatomy of a composite event detector. Technical report, CIS Department, University of Florida, December 1993.

....have to resort to programming in a powerful event control language and to influence the state variables so that no ambiguities arise. After the execution of the selected rule has terminated the events that determined its situation are discarded. In some cases this set cannot be determined uniquely [6]. In this case events are chosen indeterministically. Consider a rule fe and c: ag. There may be more than one occurrence of event e until c becomes true. 3.5 Objects Alliances enforce the collective behavior of a collection of objects. Objects are the instigators of all actions within alliances, ....

S. Chakravarthy and V. Krishnaprasad. Anatomy of a composite event detector. Technical Report UF-CIS-TR-93-039, University of Florida, Gainesville, Florida 32611, Dec. 1993.

.... as the issues of temporal databases and active rule languages are now combined and intertwined [29] In previous systems, the operational semantics of composite event detection has been defined using widely different frameworks, such as Finite State Automata [15] Petri Nets [14] or Event Graphs [4]. Even though these systems have captured many aspects of the functionality required by active database applications, they Work supported by NSF Grant IRI 9632272 This work was performed at the University of California, Los Angeles 2 IAKOVOS MOTAKIS AND CARLO ZANIOLO are restricted by the ....

....However, there are still some limitations. Simultaneous events cannot be handled. Also, it is not clear how the semantics of general negated events (as defined in the next section) can be captured by Petri nets. 4 IAKOVOS MOTAKIS AND CARLO ZANIOLO The Event Graphs framework that is used in SNOOP [4] suffers from limitations similar to those of Petri Nets. A simpler approach to overcome the mentioned problems is that of using Datalog 1S [3] This is a temporal deductive language that extends Datalog, by allowing every predicate to have (at most) one temporal argument in addition to the usual ....

S. Chakravarthy, V. Krishnaprasad, E. Anwar, and S. K. Kim. Anatomy of a Composite Event Detector. Technical Report CIS TR-93-039, University of Florida, December 1993.

....composite events that are constructed with the operators sequence, disjunction, and conjunction, whereas the latter group comprises composite events constructed with history, negation, and closure. Several different approaches are used for composite event detection including syntax graphs [Deu94, CKAK93] Petri nets [GD94] finite state automata [GJS92] and arrays [Eri93] 2.2 Conditions The condition part of a rule is usually a boolean expression, a predicate, or a set of queries, and it is satisfied if the expression evaluates to true, or all the queries return non empty results, ....

S. Chakravarthy, V. Krishnaprasad, E. Abwar, and S. K. Kim. Anatomy of a composite event detector. Technical Report UF-CIS-TR-93-039, CIS Department, University of Florida, December 1993.

....a unique terminator to link the two intervals and deliberately extend the detection interval to deal with late arriving events. 19 7 RELATED WORK AND CONCLUSIONS Event monitoring and in particular composite event detection has received much attention particularly in the area of active databases[14 17]. They provide rich notations for specifying composite events but due to their centralised nature the proposed detection algorithms are unsuitable for event composition in a loosely coupled distributed environment. In particular these algorithms assume that events are detected at the time or in ....

....expressions. The mechanism used by [20] is based on finite state machines with enhancements to allow detection of concurrent composite events. We have adopted a tree based mechanism as it seems to be the most suitable for onthe fly detection of concurrent composite events in presence of delays. [12, 14] use a similar mechanism, but our event evaluation tree maintains a hierarchical history which allows us to deal with out of order event arrivals in an efficient manner. A number of systems use a centralised correlation approach to filter and analyse events generated from distributed sources [3, ....

Chakravarthy, S., Krishnaprasad, V., Anwar, E. & Kim, S.-K. (1993) Anatomy of a Composite Event Detector, Technical Report UF-CIS-TR-93-039, University of Florida, Department of Computer and Information Sciences.

....is waiting for the action triggering event to occur. Blocking can be specified by the keyword WAIT UNTIL. In TriGS, any message sent to an object may signal a message event. Furthermore, TriGS supports time events, external events, and composite events similar to the approach described in SNOOP [Anwa94]. For each event, a guard, i.e. a predicate over the event s parameters similar to masks in ODE [Geha92] may be specified, which further restricts the events able to trigger a condition or action, respectively. The condition part of a rule is specified by a boolean expression, possibly based on ....

E. Anwar, S. Chakravarthy, S.-K. Kim, V. Krishnaprasad, Anatomy of a Composite Event Detector, Proc. of the 20th Int. Conference on Very Large Data Bases (VLDB'94), Santiago, Chile, 1994

....is waiting for the action triggering event to occur. Blocking can be specified by the keyword WAIT UNTIL. In TriGS, any message sent to an object may signal a message event. Furthermore, TriGS supports time events, external events, and composite events similar to the approach described in SNOOP [Anwa94]. For each event, a guard, i.e. a predicate over the event s parameters similar to masks in ODE [Geha92] may be specified, which further restricts the events able to trigger a condition or action, respectively. The condition part of a rule is specified by a boolean expression, possibly based on ....

E. Anwar, S. Chakravarthy, S.-K. Kim, V. Krishnaprasad, Anatomy of a Composite Event Detector, Proc. of the 20th Int. Conference on Very Large Data Bases (VLDB'94), Santiago, Chile, 1994

....contains composite events that are constructed with the operators sequence, disjunction, and conjunction, whereas the latter group comprises composite events constructed with history, negation, and closure. Several different approaches are used for composite event detection including syntax graphs [18, 11], Petri nets [22] finite state automata [24] and arrays [20] An event composition policy identifies which event occurrence of a particular event type will be used in the event composition process. Consider the CHAPTER 2. BACKGROUND 8 composite event defined as the sequence of event types ....

S. Chakravarthy, V. Krishnaprasad, E. Abwar, and S. K. Kim. Anatomy of a composite event detector. Technical Report UF-CIS-TR-93-039, CIS Department, University of Florida, December 1993.

No context found.

S. Chakravarthy, V. Krishnaprasad, E. Anwar, and S.-K. Kim. Anatomy of a composite event detector. Technical Report UF-CIS-TR-93-039, University of Florida, E470-CSE, Gainesville, FL 32611, December 1993. (Submitted for publication.).

No context found.

CHAKRAVARTHY, S., KRISHNAPRASAD, V., ANWAR, E. AND KIM, S.-K., Anatomy of a Composite Event Detector, TR-93-039, CIS, Univ. of Florida, 1993.

No context found.

S. Chakravarthy, V. Krishnaprasad. E. Anwar and S.-K Kim. Anatomy of a Composite Event Detector. UF-CIS Technical Report TR-93-039, University of Florida, 1993.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC