JBoss.orgCommunity Documentation

Chapter 2. Drools Fusion Features

2.1. Events
2.1.1. Event Semantics
2.1.2. Event Declaration
2.1.3. Event Metadata
2.2. Session Clock
2.2.1. Available Clock Implementations
2.3. Streams Support
2.3.1. Declaring and Using Entry Points
2.4. Temporal Reasoning
2.4.1. Temporal Operators
2.5. Event Processing Modes
2.5.1. Cloud Mode
2.5.2. Stream Mode
2.6. Sliding Windows
2.6.1. Sliding Time Windows
2.6.2. Sliding Length Windows
2.7. KnowledgeBase Partitioning
2.7.1. When partitioning is useful
2.7.2. How to configure partitioning
2.7.3. Multithreading management
2.8. Memory Management for Events
2.8.1. Explicit expiration offset
2.8.2. Inferred expiration offset

2.1. Events

Events, from a Drools perspective are just a special type of fact. In this way, we can say that all events are facts, but not all facts are events. In the next few sections the specific differences that characterize an event are presented.

An event is a fact that present a few distinguishing characteristics:

Drools supports the declaration and usage of events with both semantics: point-in-time events and interval-based events.

Note

A simplistic way to understand the unitification of the semantics is to consider a point-in-time event as an interval-based event whose duration is zero.

To declare a fact type as an event, all it is required is to assign the @role metadata tag to the fact type. The @role metadata tag accepts two possible values:

For instance, the example below is declaring that the fact type StockTick in a stock broker application shall be handled as an event.

Example 2.1. declaring a fact type as an event

import some.package.StockTick

declare StockTick
    @role( event )
end

The same applies to facts declared inline. So, if StockTick was a fact type declared in the DRL itself, instead of a previously existing class, the code would be:

Example 2.2. declaring a fact type and assigning it the event role

declare StockTick 
    @role( event )

    datetime : java.util.Date
    symbol : String
    price : double
end

For more information on type declarations, please check the Rule Language section of the Drools Expert documentation.

All events have a set of metadata associated to them. Most of the metadata values have defaults that are automatically assigned to each event when they are inserted into the working memory, but it is possible to change the default on an event type basis, using the metadata tags listed below.

For the examples, lets assume the user has the following class in the application domain model:

Example 2.3. the VoiceCall fact class

/**

 * A class that represents a voice call in 

 * a Telecom domain model

 */

public class VoiceCall {

    private String  originNumber;

    private String  destinationNumber;

    private Date    callDateTime;

    private long    callDuration;          // in milliseconds


    // constructors, getters and setters

}

The @role meta data was already discussed in the previous section and is presented here for completeness:

@role( <fact|event> )

It annotates a given fact type as either a regular fact or event. It accepts either "fact" or "event" as a parameter. Default is "fact".

Example 2.4. declaring VoiceCall as an event type

declare VoiceCall
    @role( event )
end

Every event has an associated timestamp assigned to it. By default, the timestamp for a given event is read from the Session Clock and assigned to the event at the time the event is inserted into the working memory. Although, sometimes, the event has the timestamp as one of it's own attributes. In this case, the user may tell the engine to use the timestamp from the event's attribute instead of reading it from the Session Clock.

@timestamp( <attributeName> )

To tell the engine what attribute to use as the source of the event's timestamp, just list the attribute name as a parameter to the @timestamp tag.

Example 2.5. declaring the VoiceCall timestamp attribute

declare VoiceCall
    @role( event )
    @timestamp( callDateTime )
end

Drools supports both event semantics: point-in-time events and interval-based events. A point-in-time event is represented as an interval-based event whose duration is zero. By default, all events have duration zero. The user may attribute a different duration for an event by declaring which attribute in the event type contains the duration of the event.

@duration( <attributeName> )

So, for our VoiceCall fact type, the declaration would be:

Example 2.6. declaring the VoiceCall duration attribute

declare VoiceCall
    @role( event )
    @timestamp( callDateTime )
    @duration( callDuration )
end

Important

This tag is only considered when running the engine in STREAM mode. Also, additional discussion on the effects of using this tag is made on the Memory Management section. It is included here for completeness.

Events may be automatically expired after some time in the working memory. Typically this happens when, based on the existing rules in the knowledge base, the event can no longer match and activate any rules. Although, it is possible to explicitly define when an event should expire.

@expires( <timeOffset> )

The value of timeOffset is a temporal interval in the form:

[#d][#h][#m][#s][#[ms]]

Where [ ] means an optional parameter and # means a numeric value.

So, to declare that the VoiceCall facts should be expired after 1 hour and 35 minutes after they are inserted into the working memory, the user would write:

Example 2.7. declaring the expiration offset for the VoiceCall events

declare VoiceCall
    @role( event )
    @timestamp( callDateTime )
    @duration( callDuration )
    @expires( 1h35m )
end

The @expires policy will take precedence and override the implicit expiration offset calculated from temporal constraints and sliding windows in the knowledge base.

2.2. Session Clock

Reasoning over time requires a reference clock. Just to mention one example, if a rule reasons over the average price of a given stock over the last 60 minutes, how the engine knows what stock price changes happened over the last 60 minutes in order to calculate the average? The obvious response is: by comparing the timestamp of the events with the "current time". How the engine knows what time is now? Again, obviously, by querying the Session Clock.

The session clock implements a strategy pattern, allowing different types of clocks to be plugged and used by the engine. This is very important because the engine may be running in an elements of different scenarios that may require different clock implementations. Just to mention a few:

Drools 5 provides 2 clock implementations out of the box. The default real time clock, based on the system clock, and an optional pseudo clock, controlled by the application.

By default, Drools uses a real time clock implementation that internally uses the system clock to determine the current timestamp.

To explicitly configure the engine to use the real time clock, just set the session configuration parameter to real time:

KnowledgeSessionConfiguration config = KnowledgeBaseFactory.newKnowledgeSessionConfiguration();

config.setOption( ClockTypeOption.get("realtime") );

Drools also offers out of the box an implementation of a clock that is controlled by the application that is called Pseudo Clock. This clock is specially useful for unit testing temporal rules since it can be controlled by the application and so the results become deterministic.

To configure the pseudo session clock, do:

KnowledgeSessionConfiguration config = KnowledgeBaseFactory.newKnowledgeSessionConfiguration();

config.setOption( ClockTypeOption.get("pseudo") );

As an example of how to control the pseudo session clock:

KnowledgeSessionConfiguration conf = KnowledgeBaseFactory.newKnowledgeSessionConfiguration();

conf.setOption( ClockTypeOption.get( "pseudo" ) );

StatefulKnowledgeSession session = kbase.newStatefulKnowledgeSession( conf, null );

        

SessionPseudoClock clock = session.getSessionClock();


// then, while inserting facts, advance the clock as necessary:

FactHandle handle1 = session.insert( tick1 );

clock.advanceTime( 10, TimeUnit.SECONDS );

FactHandle handle2 = session.insert( tick2 );

clock.advanceTime( 30, TimeUnit.SECONDS );

FactHandle handle3 = session.insert( tick3 );

2.3. Streams Support

Most CEP use cases have to deal with streams of events. The streams can be provided to the application in various forms, from JMS queues to flat text files, from database tables to raw sockets or even through web service calls. In any case, the streams share a common set of characteristics:

Drools generalized the concept of a stream as an "entry point" into the engine. An entry point is for drools a gate from which facts come. The facts may be regular facts or special facts like events.

In Drools, facts from one entry point (stream) may join with facts from any other entry point or event with facts from the working memory. Although, they never mix, i.e., they never lose the reference to the entry point through which they entered the engine. This is important because one may have the same type of facts coming into the engine through several entry points, but one fact that is inserted into the engine through entry point A will never match a pattern from a entry point B, for example.

Entry points are declared implicitly in Drools by directly making use of them in rules. I.e. referencing an entry point in a rule will make the engine, at compile time, to identify and create the proper internal structures to support that entry point.

So, for instance, lets imagine a banking application, where transactions are fed into the system coming from streams. One of the streams contains all the transactions executed in ATM machines. So, if one of the rules says: a withdraw is authorized if and only if the account balance is over the requested withdraw amount, the rule would look like:

Example 2.8. Example of Stream Usage

rule "authorize withdraw"
when
    WithdrawRequest( $ai : accountId, $am : amount ) from entry-point "ATM Stream"
    CheckingAccount( accountId == $ai, balance > $am )
then
    // authorize withdraw
end

In the previous example, the engine compiler will identify that the pattern is tied to the entry point "ATM Stream" and will both create all the necessary structures for the rulebase to support the "ATM Stream" and will only match WithdrawRequests coming from the "ATM Stream". In the previous example, the rule is also joining the event from the stream with a fact from the main working memory (CheckingAccount).

Now, lets imagine a second rule that states that a fee of $2 must be applied to any account for which a withdraw request is placed at a bank branch:

Example 2.9. Using a different Stream

rule "apply fee on withdraws on branches"
when
    WithdrawRequest( $ai : accountId, processed == true ) from entry-point "Branch Stream"
    CheckingAccount( accountId == $ai )
then
    // apply a $2 fee on the account
end

The previous rule will match events of the exact same type as the first rule (WithdrawRequest), but from two different streams, so an event inserted into "ATM Stream" will never be evaluated against the pattern on the second rule, because the rule states that it is only interested in patterns coming from the "Branch Stream".

So, entry points, besides being a proper abstraction for streams, are also a way to scope facts in the working memory, and a valuable tool for reducing cross products explosions. But that is a subject for another time.

Inserting events into an entry point is equally simple. Instead of inserting events directly into the working memory, insert them into the entry point as shown in the example below:

Example 2.10. Inserting facts into an entry point

// create your rulebase and your session as usual

StatefulKnowledgeSession session = ...


// get a reference to the entry point

WorkingMemoryEntryPoint atmStream = session.getWorkingMemoryEntryPoint( "ATM Stream" );


// and start inserting your facts into the entry point

atmStream.insert( aWithdrawRequest );

The previous example shows how to manually insert facts into a given entry point. Although, usually, the application will use one of the many adapters to plug a stream end point, like a JMS queue, directly into the engine entry point, without coding the inserts manually. The Drools pipeline API has several adapters and helpers to do that as well as examples on how to do it.

2.4. Temporal Reasoning

Temporal reasoning is another requirement of any CEP system. As discussed previously, one of the distinguishing characteristics of events is their strong temporal relationships.

Temporal reasoning is an extensive field of research, from its roots on Temporal Modal Logic to its more practical applications in business systems. There are hundreds of papers and thesis written and approaches are described for several applications. Drools once more takes a pragmatic and simple approach based on several sources, but specially worth noting the following papers:

[ALLEN81] Allen, J.F.. An Interval-based Representation of Temporal Knowledge. 1981.

[ALLEN83] Allen, J.F.. Maintaining knowledge about temporal intervals. 1983.

[BENNE00] by Bennet, Brandon and Galton, Antony P.. A Unifying Semantics for Time and Events. 2005.

[YONEK05] by Yoneki, Eiko and Bacon, Jean. Unified Semantics for Event Correlation Over Time and Space in Hybrid Network Environments. 2005.

Drools implements the Interval-based Time Event Semantics described by Allen, and represents Point-in-Time Events as Interval-based evens with duration 0 (zero).

Note

For all temporal operator intervals, the "*" (star) symbol is used to indicate positive infinity and the "-*" (minus star) is used to indicate negative infinity.

Drools implements all 13 operators defined by Allen and also their logical complement (negation). This section details each of the operators and their parameters.

The after evaluator correlates two events and matches when the temporal distance from the current event to the event being correlated belongs to the distance range declared for the operator.

Lets look at an example:

$eventA : EventA( this after[ 3m30s, 4m ] $eventB )

The previous pattern will match if and only if the temporal distance between the time when $eventB finished and the time when $eventA started is between ( 3 minutes and 30 seconds ) and ( 4 minutes ). In other words:

 3m30s <= $eventA.startTimestamp - $eventB.endTimeStamp <= 4m

The temporal distance interval for the after operator is optional:

Note

It is possible to define negative distances for this operator. Example:

$eventA : EventA( this after[ -3m30s, -2m ] $eventB )

Note

If the first value is greater than the second value, the engine automatically reverses them, as there is no reason to have the first value greater than the second value. Example: the following two patterns are considered to have the same semantics:

$eventA : EventA( this after[ -3m30s, -2m ] $eventB ) 
$eventA : EventA( this after[ -2m, -3m30s ] $eventB )

Note

The after, before and coincides operators can be used to define constraints between events, java.util.Date attributes, and long attributes (interpreted as timestamps since epoch) in any combination. Example:

EventA( this after $someDate )

The before evaluator correlates two events and matches when the temporal distance from the event being correlated to the current correlated belongs to the distance range declared for the operator.

Lets look at an example:

$eventA : EventA( this before[ 3m30s, 4m ] $eventB )

The previous pattern will match if and only if the temporal distance between the time when $eventA finished and the time when $eventB started is between ( 3 minutes and 30 seconds ) and ( 4 minutes ). In other words:

 3m30s <= $eventB.startTimestamp - $eventA.endTimeStamp <= 4m

The temporal distance interval for the before operator is optional:

Note

It is possible to define negative distances for this operator. Example:

$eventA : EventA( this before[ -3m30s, -2m ] $eventB )

Note

If the first value is greater than the second value, the engine automatically reverses them, as there is no reason to have the first value greater than the second value. Example: the following two patterns are considered to have the same semantics:

$eventA : EventA( this before[ -3m30s, -2m ] $eventB ) 
$eventA : EventA( this before[ -2m, -3m30s ] $eventB )

Note

The after, before and coincides operators can be used to define constraints between events, java.util.Date attributes, and long attributes (interpreted as timestamps since epoch) in any combination. Example:

EventA( this after $someDate )

The coincides evaluator correlates two events and matches when both happen at the same time. Optionally, the evaluator accept thresholds for the distance between events' start and finish timestamps.

Lets look at an example:

$eventA : EventA( this coincides $eventB )

The previous pattern will match if and only if the start timestamps of both $eventA and $eventB are the same AND the end timestamp of both $eventA and $eventB also are the same.

Optionally, this operator accepts one or two parameters. These parameters are the thresholds for the distance between matching timestamps.

In other words:

$eventA : EventA( this coincides[15s, 10s] $eventB )

Above pattern will match if and only if:

abs( $eventA.startTimestamp - $eventB.startTimestamp ) <= 15s && 
abs( $eventA.endTimestamp - $eventB.endTimestamp ) <= 10s

Warning

It makes no sense to use negative interval values for the parameters and the engine will raise an error if that happens.

Note

The after, before and coincides operators can be used to define constraints between events, java.util.Date attributes, and long attributes (interpreted as timestamps since epoch) in any combination. Example:

EventA( this after $someDate )

The during evaluator correlates two events and matches when the current event happens during the occurrence of the event being correlated.

Lets look at an example:

$eventA : EventA( this during $eventB )

The previous pattern will match if and only if the $eventA starts after $eventB starts and finishes before $eventB finishes.

In other words:

$eventB.startTimestamp < $eventA.startTimestamp <= $eventA.endTimestamp < $eventB.endTimestamp

The during operator accepts 1, 2 or 4 optional parameters as follow:

The finishes evaluator correlates two events and matches when the current event's start timestamp happens after the correlated event's start timestamp, but both end timestamps occur at the same time.

Lets look at an example:

$eventA : EventA( this finishes $eventB )

The previous pattern will match if and only if the $eventA starts after $eventB starts and finishes at the same time $eventB finishes.

In other words:

$eventB.startTimestamp < $eventA.startTimestamp && 
$eventA.endTimestamp == $eventB.endTimestamp

The finishes evaluator accepts one optional parameter. If it is defined, it determines the maximum distance between the end timestamp of both events in order for the operator to match. Example:

$eventA : EventA( this finishes[ 5s ] $eventB )

Will match if and only if:

$eventB.startTimestamp < $eventA.startTimestamp && 
abs( $eventA.endTimestamp - $eventB.endTimestamp ) <= 5s

Warning

It makes no sense to use a negative interval value for the parameter and the engine will raise an exception if that happens.

The finishedby evaluator correlates two events and matches when the current event start timestamp happens before the correlated event start timestamp, but both end timestamps occur at the same time. This is the symmetrical opposite of finishes evaluator.

Lets look at an example:

$eventA : EventA( this finishedby $eventB )

The previous pattern will match if and only if the $eventA starts before $eventB starts and finishes at the same time $eventB finishes.

In other words:

$eventA.startTimestamp < $eventB.startTimestamp && 
$eventA.endTimestamp == $eventB.endTimestamp

The finishedby evaluator accepts one optional parameter. If it is defined, it determines the maximum distance between the end timestamp of both events in order for the operator to match. Example:

$eventA : EventA( this finishedby[ 5s ] $eventB )

Will match if and only if:

$eventA.startTimestamp < $eventB.startTimestamp && 
abs( $eventA.endTimestamp - $eventB.endTimestamp ) <= 5s

Warning

It makes no sense to use a negative interval value for the parameter and the engine will raise an exception if that happens.

The includes evaluator correlates two events and matches when the event being correlated happens during the current event. It is the symmetrical opposite of during evaluator.

Lets look at an example:

$eventA : EventA( this includes $eventB )

The previous pattern will match if and only if the $eventB starts after $eventA starts and finishes before $eventA finishes.

In other words:

$eventA.startTimestamp < $eventB.startTimestamp <= $eventB.endTimestamp < $eventA.endTimestamp

The includes operator accepts 1, 2 or 4 optional parameters as follow:

The meets evaluator correlates two events and matches when the current event's end timestamp happens at the same time as the correlated event's start timestamp.

Lets look at an example:

$eventA : EventA( this meets $eventB )

The previous pattern will match if and only if the $eventA finishes at the same time $eventB starts.

In other words:

abs( $eventB.startTimestamp - $eventA.endTimestamp ) == 0

The meets evaluator accepts one optional parameter. If it is defined, it determines the maximum distance between the end timestamp of current event and the start timestamp of the correlated event in order for the operator to match. Example:

$eventA : EventA( this meets[ 5s ] $eventB )

Will match if and only if:

abs( $eventB.startTimestamp - $eventA.endTimestamp) <= 5s

Warning

It makes no sense to use a negative interval value for the parameter and the engine will raise an exception if that happens.

The metby evaluator correlates two events and matches when the current event's start timestamp happens at the same time as the correlated event's end timestamp.

Lets look at an example:

$eventA : EventA( this metby $eventB )

The previous pattern will match if and only if the $eventA starts at the same time $eventB finishes.

In other words:

abs( $eventA.startTimestamp - $eventB.endTimestamp ) == 0

The metby evaluator accepts one optional parameter. If it is defined, it determines the maximum distance between the end timestamp of the correlated event and the start timestamp of the current event in order for the operator to match. Example:

$eventA : EventA( this metby[ 5s ] $eventB )

Will match if and only if:

abs( $eventA.startTimestamp - $eventB.endTimestamp) <= 5s

Warning

It makes no sense to use a negative interval value for the parameter and the engine will raise an exception if that happens.

The overlaps evaluator correlates two events and matches when the current event starts before the correlated event starts and finishes after the correlated event starts, but before the correlated event finishes. In other words, both events have an overlapping period.

Lets look at an example:

$eventA : EventA( this overlaps $eventB )

The previous pattern will match if and only if:

$eventA.startTimestamp < $eventB.startTimestamp < $eventA.endTimestamp < $eventB.endTimestamp

The overlaps operator accepts 1 or 2 optional parameters as follow:

The overlappedby evaluator correlates two events and matches when the correlated event starts before the current event starts and finishes after the current event starts, but before the current event finishes. In other words, both events have an overlapping period.

Lets look at an example:

$eventA : EventA( this overlappedby $eventB )

The previous pattern will match if and only if:

$eventB.startTimestamp < $eventA.startTimestamp < $eventB.endTimestamp < $eventA.endTimestamp

The overlappedby operator accepts 1 or 2 optional parameters as follow:

The starts evaluator correlates two events and matches when the current event's end timestamp happens before the correlated event's end timestamp, but both start timestamps occur at the same time.

Lets look at an example:

$eventA : EventA( this starts $eventB )

The previous pattern will match if and only if the $eventA finishes before $eventB finishes and starts at the same time $eventB starts.

In other words:

$eventA.startTimestamp == $eventB.startTimestamp && 
$eventA.endTimestamp < $eventB.endTimestamp

The starts evaluator accepts one optional parameter. If it is defined, it determines the maximum distance between the start timestamp of both events in order for the operator to match. Example:

$eventA : EventA( this starts[ 5s ] $eventB )

Will match if and only if:

abs( $eventA.startTimestamp - $eventB.startTimestamp ) <= 5s && 
$eventA.endTimestamp < $eventB.endTimestamp

Warning

It makes no sense to use a negative interval value for the parameter and the engine will raise an exception if that happens.

The startedby evaluator correlates two events and matches when the correlating event's end timestamp happens before the current event's end timestamp, but both start timestamps occur at the same time. Lets look at an example:

$eventA : EventA( this startedby $eventB )

The previous pattern will match if and only if the $eventB finishes before $eventA finishes and starts at the same time $eventB starts.

In other words:

$eventA.startTimestamp == $eventB.startTimestamp && 
$eventA.endTimestamp > $eventB.endTimestamp

The startedby evaluator accepts one optional parameter. If it is defined, it determines the maximum distance between the start timestamp of both events in order for the operator to match. Example:

$eventA : EventA( this starts[ 5s ] $eventB )

Will match if and only if:

abs( $eventA.startTimestamp - $eventB.startTimestamp ) <= 5s && 
$eventA.endTimestamp > $eventB.endTimestamp

Warning

It makes no sense to use a negative interval value for the parameter and the engine will raise an exception if that happens.

2.5. Event Processing Modes

Rules engines in general have a well known way of processing data and rules and provide the application with the results. Also, there is not many requirements on how facts should be presented to the rules engine, specially because in general, the processing itself is time independent. That is a good assumption for most scenarios, but not for all of them. When the requirements include the processing of real time or near real time events, time becomes and important variable of the reasoning process.

The following sections will explain the impact of time on rules reasoning and the two modes provided by Drools for the reasoning process.

The CLOUD processing mode is the default processing mode. Users of rules engine are familiar with this mode because it behaves in exactly the same way as any pure forward chaining rules engine, including previous versions of Drools.

When running in CLOUD mode, the engine sees all facts in the working memory, does not matter if they are regular facts or events, as a whole. There is no notion of flow of time, although events have a timestamp as usual. In other words, although the engine knows that a given event was created, for instance, on January 1st 2009, at 09:35:40.767, it is not possible for the engine to determine how "old" the event is, because there is no concept of "now".

In this mode, the engine will apply its usual many-to-many pattern matching algorithm, using the rules constraints to find the matching tuples, activate and fire rules as usual.

This mode does not impose any kind of additional requirements on facts. So for instance:

On the other hand, since there is no requirements, some benefits are not available either. For instance, in CLOUD mode, it is not possible to use sliding windows, because sliding windows are based on the concept of "now" and there is no concept of "now" in CLOUD mode.

Since there is no ordering requirement on events, it is not possible for the engine to determine when events can no longer match and as so, there is no automatic life-cycle management for events. I.e., the application must explicitly retract events when they are no longer necessary, in the same way the application does with regular facts.

Cloud mode is the default execution mode for Drools, but in any case, as any other configuration in Drools, it is possible to change this behavior either by setting a system property, using configuration property files or using the API. The corresponding property is:

KnowledgeBaseConfiguration config = KnowledgeBaseFactory.newKnowledgeBaseConfiguration();

config.setOption( EventProcessingOption.CLOUD );

The equivalent property is:

drools.eventProcessingMode = cloud

The STREAM processing mode is the mode of choice when the application needs to process streams of events. It adds a few common requirements to the regular processing, but enables a whole lot of features that make stream event processing a lot simpler.

The main requirements to use STREAM mode are:

Given that the above requirements are met, the application may enable the STREAM mode using the following API:

KnowledgeBaseConfiguration config = KnowledgeBaseFactory.newKnowledgeBaseConfiguration();

config.setOption( EventProcessingOption.STREAM );

Or, the equivalent property: 

drools.eventProcessingMode = stream

When using the STREAM, the engine knows the concept of flow of time and the concept of "now", i.e., the engine understands how old events are based on the current timestamp read from the Session Clock. This characteristic allows the engine to provide the following additional features to the application:

All these features are explained in the following sections.

When running the engine in CLOUD mode, the session clock is used only to time stamp the arriving events that don't have a previously defined timestamp attribute. Although, in STREAM mode, the Session Clock assumes an even more important role.

In STREAM mode, the session clock is responsible for keeping the current timestamp, and based on it, the engine does all the temporal calculations on event's aging, synchronizes streams from multiple sources, schedules future tasks and so on.

Check the documentation on the Session Clock section to know how to configure and use different session clock implementations.

Negative patterns behave different in STREAM mode when compared to CLOUD mode. In CLOUD mode, the engine assumes that all facts and events are known in advance (there is no concept of flow of time) and so, negative patterns are evaluated immediately.

When running in STREAM mode, negative patterns with temporal constraints may require the engine to wait for a time period before activating a rule. The time period is automatically calculated by the engine in a way that the user does not need to use any tricks to achieve the desired result.

For instance:

Example 2.11. a rule that activates immediately upon matching

rule "Sound the alarm"
when
    $f : FireDetected( )
    not( SprinklerActivated( ) )
then
    // sound the alarm
end

The above rule has no temporal constraints that would require delaying the rule, and so, the rule activates immediately. The following rule on the other hand, must wait for 10 seconds before activating, since it may take up to 10 seconds for the sprinklers to activate:

Example 2.12. a rule that automatically delays activation due to temporal constraints

rule "Sound the alarm"
when
    $f : FireDetected( )
    not( SprinklerActivated( this after[0s,10s] $f ) )
then
    // sound the alarm
end

This behaviour allows the engine to keep consistency when dealing with negative patterns and temporal constraints at the same time. The above would be the same as writing the rule as below, but does not burden the user to calculate and explicitly write the appropriate duration parameter:

Example 2.13. same rule with explicit duration parameter

rule "Sound the alarm"
    duration( 10s )
when
    $f : FireDetected( )
    not( SprinklerActivated( this after[0s,10s] $f ) )
then
    // sound the alarm
end

The following rule expects every 10 seconds at least one “Heartbeat” event, if not the rule fires. The special case in this rule is that we use the same type of the object in the first pattern and in the negative pattern. The negative pattern has the temporal constraint to wait between 0 to 10 seconds before firing and it excludes the Heartbeat bound to $h. Excluding the bound Heartbeat is important since the temporal constraint [0s, ...] does not exclude by itself the bound event $h from being matched again, thus preventing the rule to fire.

Example 2.14. excluding bound events in negative patterns

rule "Sound the alarm"
when
    $h: Heartbeat( ) from entry-point "MonitoringStream"
    not( Heartbeat( this != $h, this after[0s,10s] $h ) from entry-point "MonitoringStream" )
then
    // Sound the alarm
end

2.6. Sliding Windows

Sliding Windows are a way to scope the events of interest by defining a window that is constantly moving. The two most common types of sliding window implementations are time based windows and length based windows.

The next sections will detail each of them.

Important

Sliding Windows are only available when running the engine in STREAM mode. Check the Event Processing Mode section for details on how the STREAM mode works.

Important

Sliding windows start to match immediately and defining a sliding window does not imply that the rule has to wait for the sliding window to be "full" in order to match. For instance, a rule that calculates the average of an event property on a window:length(10) will start calculating the average immediately, and it will start at 0 (zero) for no-events, and will update the average as events arrive one by one.

Sliding Time Windows allow the user to write rules that will only match events occurring in the last X time units.

For instance, if the user wants to consider only the Stock Ticks that happened in the last 2 minutes, the pattern would look like this:

StockTick() over window:time( 2m )

Drools uses the "over" keyword to associate windows to patterns.

On a more elaborate example, if the user wants to sound an alarm in case the average temperature over the last 10 minutes read from a sensor is above the threshold value, the rule would look like:

Example 2.15. aggregating values over time windows

rule "Sound the alarm in case temperature rises above threshold"
when
    TemperatureThreshold( $max : max )
    Number( doubleValue > $max ) from accumulate(
        SensorReading( $temp : temperature ) over window:time( 10m ),
        average( $temp ) )
then
    // sound the alarm
end

The engine will automatically disregard any SensorReading older than 10 minutes and keep the calculated average consistent.

Important

Please note that time based windows are considered when calculating the interval an event remains in the working memory before being expired, but an event falling off a sliding window does not mean by itself that the event will be discarded from the working memory, as there might be other rules that depend on that event. The engine will discard events only when no other rules depend on that event and the expiration policy for that event type is fulfilled.

Sliding Length Windows work the same way as Time Windows, but consider events based on order of their insertion into the session instead of flow of time.

For instance, if the user wants to consider only the last 10 RHT Stock Ticks, independent of how old they are, the pattern would look like this:

StockTick( company == "RHT" ) over window:length( 10 )

As you can see, the pattern is similar to the one presented in the previous section, but instead of using window:time to define the sliding window, it uses window:length.

Using a similar example to the one in the previous section, if the user wants to sound an alarm in case the average temperature over the last 100 readings from a sensor is above the threshold value, the rule would look like:

Example 2.16. aggregating values over length windows

rule "Sound the alarm in case temperature rises above threshold"
when
    TemperatureThreshold( $max : max )
    Number( doubleValue > $max ) from accumulate(
        SensorReading( $temp : temperature ) over window:length( 100 ),
        average( $temp ) )
then
    // sound the alarm
end

The engine will keep only consider the last 100 readings to calculate the average temperature.

Important

Please note that falling off a length based window is not criteria for event expiration in the session. The engine disregards events that fall off a window when calculating that window, but does not remove the event from the session based on that condition alone as there might be other rules that depend on that event.

Important

Please note that length based windows do not define temporal constraints for event expiration from the session, and the engine will not consider them. If events have no other rules defining temporal constraints and no explicit expiration policy, the engine will keep them in the session indefinitely.

2.7. KnowledgeBase Partitioning

Warning

THIS FEATURE IS NO LONGER SUPPORTED. It may be reintroduced in the future, but trying to use it in the current version will raise an exception.

The classic Rete algorithm is usually executed using a single thread. Although, as confirmed in several opportunities by Dr. Forgy, the algorithm itself is parallelizable. Drools implementation of the ReteOO algorithm supports coarse grained parallelization through rulebase partitioning.

When this option is enabled, the rulebase will be partitioned in several independent partitions and a pool of worker threads will be used to propagate facts through the partitions. The implementation guarantees that at most one worker thread will be executing tasks for a given partition, but multiple partitions may be "active" at a single point in time.

Everything should be transparent to the user, except that all working memory actions (insert/retract/modify) are executed asynchronously.

Important

This feature enables parallel LHS evaluation, but does not change the behavior of rule firing. I.e., rules will continue to fire sequentially, according to the conflict resolution strategy.

Knowledge base partitioning is a very powerful feature for specific scenarios, but it is not a general case solution. To understand if this feature would be useful for a given scenario, the user may follow the checklist below:

  1. Does your hardware contains multiple processors?

  2. Does your knowledge session process a high volume of facts?

  3. Are the LHS of your rules expensive to evaluate? (ex: use expensive "from" expressions)

  4. Does your knowledge base contains hundreds or more rules?

If the answer to all the questions above is "yes", then this feature will probably increase the overall performance of your rulebase evaluation.

To enable knowledge base partitioning, set the following option:

Example 2.17. Enabling multithread evaluation (partitioning)

KnowledgeBaseConfiguration config = KnowledgeBaseFactory.newKnowledgeBaseConfiguration();

config.setOption( MultithreadEvaluationOption.YES );

The equivalent property is:

drools.multithreadEvaluation = <true|false>

The default value for this option is "false" (disabled).

Drools offers a simple configuration option for users to control the size of the worker thread's pool.

To define the maximum size for the thread pool, the user may use the following configuration option:

Example 2.18. setting the maximum number of threads for rule evaluation to 5

KnowledgeBaseConfiguration config = KnowledgeBaseFactory.newKnowledgeBaseConfiguration();

config.setOption( MaxThreadsOption.get(5) );

The equivalent property is:

drools.maxThreads = <-1|1..n>

The default value for this configuration is 3 and a negative number means the engine will try to spawn as many threads as there are partitions in the rulebase.

Warning

It is usually dangerous to set this option with a negative number. Always set it with a sensible positive number of threads.

2.8. Memory Management for Events

Important

The automatic memory management for events is only performed when running the engine in STREAM mode. Check the Event Processing Mode section for details on how the STREAM mode works.

One of the benefits of running the engine in STREAM mode is that the engine can detect when an event can no longer match any rule due to its temporal constraints. When that happens, the engine can safely retract the event from the session without side effects and release any resources used by that event.

There are basically 2 ways for the engine to calculate the matching window for a given event:

The first way of allowing the engine to calculate the window of interest for a given event type is by explicitly setting it. To do that, just use the declare statement and define an expiration for the fact type:

Example 2.19. explicitly defining an expiration offset of 30 minutes for StockTick events

declare StockTick
    @expires( 30m )
end

The above example declares an expiration offset of 30 minutes for StockTick events. After that time, assuming no rule still needs the event, the engine will expire and remove the event from the session automatically.

Another way for the engine to calculate the expiration offset for a given event is implicitly, by analyzing the temporal constraints in the rules. For instance, given the following rule:

Example 2.20. example rule with temporal constraints

rule "correlate orders"
when
    $bo : BuyOrderEvent( $id : id ) 
    $ae : AckEvent( id == $id, this after[0,10s] $bo )
then
    // do something
end

Analyzing the above rule, the engine automatically calculates that whenever a BuyOrderEvent matches, it needs to store it for up to 10 seconds to wait for matching AckEvent's. So, the implicit expiration offset for BuyOrderEvent will be 10 seconds. AckEvent, on the other hand, can only match existing BuyOrderEvent's, and so its expiration offset will be zero seconds.

The engine will make this analysis for the whole rulebase and find the offset for every event type. Whenever an implicit expiration offset clashes with the explicit expiration offset, then engine will use the greater of the two.