com.sun.slamd.job
Class SingleStatisticWithReplicationLatencyOptimizationAlgorithm

java.lang.Object
  com.sun.slamd.job.OptimizationAlgorithm
      com.sun.slamd.job.SingleStatisticWithReplicationLatencyOptimizationAlgorithm

public class SingleStatisticWithReplicationLatencyOptimizationAlgorithm
extends OptimizationAlgorithm

This class defines a SLAMD optimization algorithm that looks at a single statistic within the job and finds the iteration with the highest or lowest value for a given stat tracker that also has acceptable LDAP replication latency characteristics. In particular, it can reject any iteration with an average latency greater than a given value, and it can also reject any iteration where the average of the last 25% of the iterations is greater than the average of the first 25% of the iterations by a specified percentage.

For this optimization algorithm to be used, an optimizing job must include the appropriate data from at least one replication latency resource monitor. If there are multiple replication latency trackers used for the optimizing job, then they will all be taken into consideration.

Author:

Neil A. Wilson

Field Summary

static int	OPTIMIZE_TYPE_MAXIMIZE The optimization type value that indicates that we should try to find the highest value for the statistic to optimize.
static int	OPTIMIZE_TYPE_MINIMIZE The optimization type value that indicates that we should try to find the lowest value for the statistic to optimize.
static java.lang.String	PARAM_MAX_PERCENT_INCREASE The name of the parameter that is used to specify the maximum allowed percentage of increase in latency between the beginning of the job and the end of the job.
static java.lang.String	PARAM_MAX_REPLICA_LATENCY The name of the parameter that is used to specify the maximum replication latency that will be acceptable.
static java.lang.String	PARAM_MIN_PCT_IMPROVEMENT The name of the parameter that is used to specify the minimum required percent improvement needed for a new best iteration.
static java.lang.String	PARAM_OPTIMIZE_STAT The name of the parameter that is used to specify the statistic to optimize.
static java.lang.String	PARAM_OPTIMIZE_TYPE The name of the parameter that is used to specify the type of optimization to perform.

Constructor Summary

SingleStatisticWithReplicationLatencyOptimizationAlgorithm()
Creates a new instance of this optimization algorithm.

Method Summary

boolean	availableWithJobClass(JobClass jobClass) Indicates whether this optimization algorithm may be used when running the specified type of job.
double	getIterationOptimizationValue(Job iteration) Retrieves the value associated with the provided iteration of the given optimizing job.
java.lang.String	getOptimizationAlgorithmName() Retrieves the human-readable name that will be used for this optimization algorithm.
ParameterList	getOptimizationAlgorithmParameters() Retrieves the set of parameters that have been defined for this optimization algorithm.
ParameterList	getOptimizationAlgorithmParameterStubs(JobClass jobClass) Retrieves a set of parameter stubs that should be used to prompt the end user for the settings to use when executing the optimizing job.
void	initializeOptimizationAlgorithm(OptimizingJob optimizingJob, ParameterList parameters) Initializes this optimization algorithm with the provided set of parameters for the given optimizing job.
boolean	isBestIterationSoFar(Job iteration) Indicates whether the provided iteration is the best one seen so far for the given optimizing job based on the constraints specified in the parameters used to initialize this optimization algorithm.
OptimizationAlgorithm	newInstance() Creates a new, uninitialized instance of this optimization algorithm.
void	reInitializeOptimizationAlgorithm() Clears any state information currently set for this optimization algorithm and restores it to the state it would have if a new instance had been created and only the initializeOptimizationAlgorithm() method had been called on that instance.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

PARAM_MIN_PCT_IMPROVEMENT

public static final java.lang.String PARAM_MIN_PCT_IMPROVEMENT

The name of the parameter that is used to specify the minimum required percent improvement needed for a new best iteration.

See Also:

Constant Field Values

PARAM_OPTIMIZE_STAT

public static final java.lang.String PARAM_OPTIMIZE_STAT

The name of the parameter that is used to specify the statistic to optimize.

See Also:

Constant Field Values

PARAM_OPTIMIZE_TYPE

public static final java.lang.String PARAM_OPTIMIZE_TYPE

The name of the parameter that is used to specify the type of optimization to perform.

See Also:

Constant Field Values

PARAM_MAX_REPLICA_LATENCY

public static final java.lang.String PARAM_MAX_REPLICA_LATENCY

The name of the parameter that is used to specify the maximum replication latency that will be acceptable.

See Also:

Constant Field Values

PARAM_MAX_PERCENT_INCREASE

public static final java.lang.String PARAM_MAX_PERCENT_INCREASE

The name of the parameter that is used to specify the maximum allowed percentage of increase in latency between the beginning of the job and the end of the job.

See Also:

Constant Field Values

OPTIMIZE_TYPE_MAXIMIZE

public static final int OPTIMIZE_TYPE_MAXIMIZE

The optimization type value that indicates that we should try to find the highest value for the statistic to optimize.

See Also:

Constant Field Values

OPTIMIZE_TYPE_MINIMIZE

public static final int OPTIMIZE_TYPE_MINIMIZE

The optimization type value that indicates that we should try to find the lowest value for the statistic to optimize.

See Also:

Constant Field Values

Constructor Detail

SingleStatisticWithReplicationLatencyOptimizationAlgorithm

public SingleStatisticWithReplicationLatencyOptimizationAlgorithm()

Creates a new instance of this optimization algorithm. All subclasses must define a constructor that does not take any arguments, and they must invoke the constructor of this superclass as their first action.

Method Detail

getOptimizationAlgorithmName

public java.lang.String getOptimizationAlgorithmName()

Retrieves the human-readable name that will be used for this optimization algorithm.

Specified by:

getOptimizationAlgorithmName in class OptimizationAlgorithm

Returns:

The human-readable name that will be used for this optimization algorithm.

newInstance

public OptimizationAlgorithm newInstance()

Creates a new, uninitialized instance of this optimization algorithm. In most cases, this should simply return the object created from invoking the default constructor.

Specified by:

newInstance in class OptimizationAlgorithm

Returns:

The new instance of this optimization algorithm.

availableWithJobClass

public boolean availableWithJobClass(JobClass jobClass)

Indicates whether this optimization algorithm may be used when running the specified type of job. This algorithm is only available for jobs that report at least one "searchable" stat tracker.

Specified by:

availableWithJobClass in class OptimizationAlgorithm

Parameters:

jobClass - The job class for which to make the determination.

Returns:

true if this optimization algorithm may be used with the provided job class, or false if not.

reInitializeOptimizationAlgorithm

public void reInitializeOptimizationAlgorithm()

Clears any state information currently set for this optimization algorithm and restores it to the state it would have if a new instance had been created and only the initializeOptimizationAlgorithm() method had been called on that instance. In this case, all that is necessary is to forget about the best value seen so far.

Specified by:

reInitializeOptimizationAlgorithm in class OptimizationAlgorithm

getOptimizationAlgorithmParameterStubs

public ParameterList getOptimizationAlgorithmParameterStubs(JobClass jobClass)

Retrieves a set of parameter stubs that should be used to prompt the end user for the settings to use when executing the optimizing job.

Specified by:

getOptimizationAlgorithmParameterStubs in class OptimizationAlgorithm

Parameters:

jobClass - The job class that will be used for the optimizing job.

Returns:

A set of parameter stubs that should be used to prompt the end user for the settings to use when executing the optimizing job.

getOptimizationAlgorithmParameters

public ParameterList getOptimizationAlgorithmParameters()

Retrieves the set of parameters that have been defined for this optimization algorithm.

Specified by:

getOptimizationAlgorithmParameters in class OptimizationAlgorithm

Returns:

The set of parameters that have been defined for this optimization algorithm.

initializeOptimizationAlgorithm

public void initializeOptimizationAlgorithm(OptimizingJob optimizingJob,
                                            ParameterList parameters)
                                     throws InvalidValueException

Initializes this optimization algorithm with the provided set of parameters for the given optimizing job.

Specified by:

initializeOptimizationAlgorithm in class OptimizationAlgorithm

Parameters:

optimizingJob - The optimizing job with which this optimization algorithm will be used.

parameters - The parameter list containing the parameter values provided by the end user when scheduling the optimizing job.

Throws:

InvalidValueException - If the contents of the provided parameter list are not valid for use with this optimization algorithm.

isBestIterationSoFar

public boolean isBestIterationSoFar(Job iteration)
                             throws SLAMDException

Indicates whether the provided iteration is the best one seen so far for the given optimizing job based on the constraints specified in the parameters used to initialize this optimization algorithm.

Specified by:

isBestIterationSoFar in class OptimizationAlgorithm

Parameters:

iteration - The job iteration for which to make the determination.

Returns:

true if the provided iteration is the best one seen so far for the optimizing job, or false if not.

Throws:

SLAMDException - If a problem occurs that prevents a valid determination from being made. If this exception is thrown, then the optimizing job will stop immediately with no further iterations.

getIterationOptimizationValue

public double getIterationOptimizationValue(Job iteration)
                                     throws SLAMDException

Retrieves the value associated with the provided iteration of the given optimizing job.

Specified by:

getIterationOptimizationValue in class OptimizationAlgorithm

Parameters:

iteration - The job iteration for which to retrieve the value.

Returns:

The value associated with the provided iteration of the given optimizing job.

Throws:

SLAMDException - If a problem occurs while trying to determine the value for the given optimizing job iteration.