Discrete Event Simulation of Complex Systems with DEUS Marco Picone - - PowerPoint PPT Presentation

▶
discrete event simulation of complex systems with deus
SMART_READER_LITE
LIVE PREVIEW

Discrete Event Simulation of Complex Systems with DEUS Marco Picone - - PowerPoint PPT Presentation

Oct 24, 2022 582 likes •823 views

Universit Degli Studi di Parma D istributed S ystems G roup Discrete Event Simulation of Complex Systems with DEUS Marco Picone , Michele Amoretti, Gianni Conte InfQ 2012 - 5/6 July - Lucca, Italy Universit Degli Studi di Parma D istributed

slide-1
SLIDE 1

Distributed Systems Group Università Degli Studi di Parma

InfQ 2012 - 5/6 July - Lucca, Italy

Discrete Event Simulation of Complex Systems with DEUS

Marco Picone, Michele Amoretti, Gianni Conte

slide-2
SLIDE 2

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

  • DEUS is multi-platform discrete event simulator, being developed with Java language.
  • It is a Open Source project available on Google Code repository (http://code.google.com/p/

deus/).

  • Has been used to simulate:
  • Distributed / Peer-to-Peer Systems
  • Vehicular Networks
  • Social Networks
  • Autonomic & Cloud Computing
  • Smart Grid

2

  • M. Amoretti, M. Agosti, F. Zanichelli, DEUS: a Discrete Event Universal Simulator, (SIMUTools 2009), Roma, Italy, March 2009.
slide-3
SLIDE 3

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

DEUS

  • Its API allows developers to implement (by subclassing)
  • nodes: the parts which interact in a complex system, leading to

emergent behaviors: humans, pets, cells, robots, intelligent agents, vehicles etc.;

  • events: e.g. node births/deaths, interactions among nodes,

interactions with the environment, logs, etc.;

  • processes: either stochastic or deterministic ones, constraining the

timeliness of events.

3

slide-4
SLIDE 4

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

DEUS

  • The simulation behavior follows the standard model of discrete event

simulations:

  • initialization of system state variables and clock
  • scheduling of initial events and, until the ending condition is true
  • calculation of next clock time
  • processing of the next event in the scheduling queue.
  • The simulation configuration is defined using an XML file containing

declaration of nodes, events and processes and the relationships between them.

4

slide-5
SLIDE 5

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

Core Packages

  • Event: the simulation object being scheduled by the Engine
  • Node: a generic data structure collector
  • Resource: a generic resource associated to a node
  • Process: the simulation object responsible of the timing of event

scheduling)

  • Engine : DEUS simulation engine
  • AutomatorParser: manages the simulation configuration file according to

DEUS XML Schema

5

slide-6
SLIDE 6

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

P2P Extension

  • p2p.event subpackage
  • SingleConnectionEvent
  • MultipleRandomConnectionEvent
  • ExpTopologyConnectionEvent
  • DisconnectionEvent
  • LogNodeDegreeEvent
  • p2p.node subpackage
  • Peer (key, list of neighbors and connection status)

6

slide-7
SLIDE 7

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

Visual Editor

7

slide-8
SLIDE 8

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

DEUS AUTOMATOR

8

slide-9
SLIDE 9

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

Simulation Automator

9

DEUS

Automator DEUS Engine

1 N

Automator GUI

Parameters Step Initial Value Final Value

Aggregated Results [Average, STDEV, ...]

Seed

slide-10
SLIDE 10

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012 10

  • Smart cities are envisioned to generate and consume overwhelming amount of data which can be

harnessed to provide relevant information about their status so as to enhance the security and lifestyle of their citizens.

  • Location-Based Service (LBS) is an information or entertainment service, accessible with mobile

devices through the mobile network and utilizing the ability to make use of the geographical position of the mobile device.

  • Discrete event simulation is a powerful means to aid the design of the enabling ICT infrastructure

for smart cities and generally location based services, in particular as a tool to predict the impact

  • n user behaviors to the purpose of improving key urban business processes.
slide-11
SLIDE 11

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

Mobility Model

  • Mobility models (MM) represent the movement of mobile users, and how their location, velocity and

acceleration change over time.

  • MM are crucial for example to test and evaluate vehicle-to-vehicle (V2V) or vehicle-to-infrastructure (V2I)
  • r Smart Cities networking applications in realistic scenario.
  • Principal characteristics are:
  • Accurate and realistic topological maps
  • Obstacles
  • Node characteristics
  • Trip and Path motion
  • Human driving/moving patterns
  • External influences
  • Intersection management

11

slide-12
SLIDE 12

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

Virtual Track & Switch Station Model

12

S1 S3 S2 S5 S4

  • B. Zhou, K. Xu, M. Gerla. Group and swarm mobility models for ad hoc network scenarios using virtual tracks, IEEE Military Communications Conference (MILCOM) 2004.
slide-13
SLIDE 13

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

Virtual Track & Switch Station Model

  • First of all the developer need to define the

switch stations and virtual tracks.

  • Switch Stations could be randomly located in

the region of interest or their geographic location could be decided by the designer.

  • A single node or a group of nodes are initially

located on a single Switch Station or along a selected Virtual Track.

  • At the switch station, the group or the single car

can randomly select one track from all other tracks at that station for the next movement.

13

slide-14
SLIDE 14

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

DEUS & Mobility

  • IMobilePeer: Interface of DEUS mobile peer
  • MobilePeer: Provides fundamental implementation for a MobilePeer based on Virtual Tracks

and Switch Station.

  • IMobilityModel: Interface for mobility model.
  • GeoLocation: Represents geographic point with its latitude and longitude.
  • SwitchStation & MobilityPathPoint: Extensions of GeoLocation.
  • MobilityPath: Describes a virtual track with its points and specific characteristics.
  • MobilityPathIndex: Keeps information about node position on a specific path.
  • MovePeerEvent: DEUS event used to simulate the move of a peer along a path.
  • SwitchStationController: Load switch stations and paths from input files.

14

slide-15
SLIDE 15

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

Mobile Peer

  • MobilePeer is an extension of Node class of DEUS. It provides a dedicated API with all

necessary classes and methods to support mobility.

  • In detail it has all methods and required callbacks to :
  • Initialize and configure the Mobility Model
  • Start or Stop node movements
  • Schedule the new movement
  • Notify the developer when the node
  • changes its geographic location
  • reaches the end of the current path and should select the new one according to its state or

to environment conditions

15

slide-16
SLIDE 16

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

slide-17
SLIDE 17

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012 17

slide-18
SLIDE 18

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012 18

slide-19
SLIDE 19

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012 19

slide-20
SLIDE 20

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012 20

slide-21
SLIDE 21

Distributed Systems Group Università Degli Studi di Parma

Marco Picone - 2012

Conclusion and Future Work

  • DEUS has been successfully used to design and simulate several and heterogeneous scenarios

V2V and V2I application, Vertical Handover Analysis and P2P & Network Coding Approach for Information Management in Smart Cities.

  • Simulation on a single machine reaches a linear density up to 40 Peer/Km and allowed to easily

customize the simulation and the associated graphical visualization tool to easily understand the behavior of active nodes.

  • Future Activities
  • New DEUS Automator, Examples and Tutorials
  • Release the firs public version of Distributed DEUS
  • Integration of Open Street Map to retrieve detailed virtual tracks points and to build an offline

monitoring tool

  • Complete and release the first version 3D visualization module.
  • Introduce the possibility to simulate and visualize mobility of nodes in indoor scenarios.

21

slide-22
SLIDE 22

Distributed Systems Group Università Degli Studi di Parma

InfQ 2012 - 5/6 July - Lucca, Italy

Marco Picone

http://dsg.ce.unipr.it/picone http://www.mendeley.com/profiles/marco-picone/