MarkovScheduler

simple module

This module implements a packet scheduler using a Markov process that has as many states as input gates the scheduler has. The input gate for the next packet is determined by the current state of the Markov process.

Inheritance diagram

The following diagram shows inheritance relationships for this type. Unresolved types are missing from the diagram.

Used in compound modules

Name	Type	Description
BurstyPacketProducer	compound module	This module mixes two different packet sources to generate bursty traffic.

Extends

Name	Type	Description
PacketSchedulerBase	simple module	This is a base module for various packet scheduler modules. Derived modules must implement a single packet scheduler function which determines the index of the input gate for the pulled packet.

Parameters

Name	Type	Default value	Description
displayStringTextFormat	string	"classified %p pk (%l)\ncurrent state: %s"	determines the text that is written on top of the submodule
reverseOrder	bool	false
clockModule	string	""	relative path of a module that implements IClock; optional
initialState	int	0	the start state of the Markov process
transitionProbabilities	string		the transition matrix (N x N) of the Markov process, specified as a list of probabilities
waitIntervals	string		the amount of time the Markov process stays in a given state, a list of intervals (N) indexed by the state

Properties

Name	Value	Description
display	i=block/join
class	MarkovScheduler

Gates

Name	Direction	Size	Description
in [ ]	input
out	output

Signals

Name	Type	Unit
packetPulled	inet::Packet

Statistics

Name	Title	Source	Record	Unit	Interpolation Mode
packetLengths	packet lengths	packetLength(packetPulled)	sum, histogram, vector	b	none
dataRate	data rate	throughput(packetPulled)	vector	bps	linear
packets	packets	packetPulled	count	pk

Scheduled messages (observed)

msg	kind	ctrl	tags	msgname	context
ClockEvent	0			WaitTimer

Direct method calls (observed)

call to	function	info
PacketCloner	inet::queueing::PacketCloner::pushPacket	pushPacket
ActivePacketSink	inet::queueing::ActivePacketSink::handleCanPullPacketChanged	handleCanPullPacketChanged
ActivePacketSource	inet::queueing::ActivePacketSource::handleCanPushPacketChanged	handleCanPushPacketChanged
EmptyPacketSource	inet::queueing::EmptyPacketSource::handleCanPushPacketChanged	handleCanPushPacketChanged
PassivePacketSource	inet::queueing::PassivePacketSource::pullPacket	pullPacket

Called methods (observed)

function	info	call from
inet::queueing::MarkovScheduler::handleCanPullPacketChanged	handleCanPullPacketChanged	PassivePacketSource
inet::queueing::MarkovScheduler::handleCanPushPacketChanged	handleCanPushPacketChanged	PacketCloner
inet::queueing::MarkovScheduler::pullPacket	pullPacket	ActivePacketSink
inet::queueing::MarkovScheduler::pushPacket	pushPacket	ActivePacketSource

Outgoing messages (observed)

gate	msg	kind	ctrl	destModule	tags
out	Packet	0		TcpServerSocketIo	DirectionTag
out	Packet	0		UdpSocketIo	DirectionTag

Pull in messages (observed)

gate	msg	kind	ctrl	srcModule	tags
in[]	Packet	0		PassivePacketSource	DirectionTag

Pulled out messages (observed)

gate	msg	kind	ctrl	destModule	tags
out	Packet	0		ActivePacketSink	DirectionTag

Pushed in messages (observed)

gate	msg	kind	ctrl	srcModule	tags
in[]	Packet	0		ActivePacketSource	DirectionTag

Push out messages (observed)

gate	msg	kind	ctrl	destModule	tags
out	Packet	0		PacketCloner	DirectionTag

Source code

//
// This module implements a packet scheduler using a Markov process that has
// as many states as input gates the scheduler has. The input gate for the next
// packet is determined by the current state of the Markov process.
//
simple MarkovScheduler extends PacketSchedulerBase like IPacketScheduler
{
    parameters:
        displayStringTextFormat = default("classified %p pk (%l)\ncurrent state: %s");
        string clockModule = default(""); // relative path of a module that implements IClock; optional
        int initialState = default(0); // the start state of the Markov process
        string transitionProbabilities; // the transition matrix (N x N) of the Markov process, specified as a list of probabilities
        string waitIntervals; // the amount of time the Markov process stays in a given state, a list of intervals (N) indexed by the state
        @class(MarkovScheduler);
}

File: src/inet/queueing/scheduler/MarkovScheduler.ned