+ All Categories
Home > Documents > 24 Lecture Wosinska

24 Lecture Wosinska

Date post: 05-Apr-2018
Category:
Upload: alextm76
View: 215 times
Download: 0 times
Share this document with a friend

of 34

Transcript
  • 8/2/2019 24 Lecture Wosinska

    1/34

    9/24/2009

    1

    Photonics in Switching

    Lena Wosinska, Assoc. Professor

    Royal Institute of Technology KTH

    Electrum 229, 164 40 Kista, Sweden

    [email protected]

    The aim of this tutorial

    To s ow princip es or optica circuitswitching and optical packet switching

    To highlight the main technologicalproblems

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 2

    To give some examples of the opticalswitching node architectures

  • 8/2/2019 24 Lecture Wosinska

    2/34

    9/24/2009

    2

    Outline

    Introduction

    Photonic Circuit Switching WDM network design

    WDM network elements

    Photonic Packet Switching

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    3

    Contention resolution

    Summary

    How to connect end users ?

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    4

  • 8/2/2019 24 Lecture Wosinska

    3/34

    9/24/2009

    3

    Switched networks

    Switching nodes not concerned with contents

    of data purpose: provide switching

    facility in general not fully connected

    End nodes provides data to transfer

    1

    A

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    5

    nodesLinks

    physical connections betweennodes

    Switching

    Circuit

    Switching

    Packet

    Switching

    Switching

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    6

    Connection Oriented

    (Virtual Circuit)

    Datagram

  • 8/2/2019 24 Lecture Wosinska

    4/34

    9/24/2009

    4

    Development ofoptical networks

    First-generation optical networks

    Transmission in the optical domain (to provide capacity)

    Example: SONET network (Synchronous Optical Network)

    Second-generation optical networks

    More functionality in the optical domain (optical networking)

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 7

    Some of routing, switching and intelligence is moving into theoptical layer

    Third-generation optical networks (?)

    SDH/SONET Networking

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    8

  • 8/2/2019 24 Lecture Wosinska

    5/34

    9/24/2009

    5

    Photonics in switching

    O tical circuit switchin OCS Wavelength-routed networks

    Relatively mature technology today

    Providing lightpaths

    WDM network elements: OLT, OADM, OXC

    (All)optical packet switching (OPS) Not available today due to some technological problems

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    9

    on ro a e op ca memory or op ca u er ng Control functions in the optical domain

    Synchronization, etc

    Optical burst switching (OBS) A feasible solution?

    Optical Circuit Switching

    W ave leng th -Rou t ed Ne tw o rk s

  • 8/2/2019 24 Lecture Wosinska

    6/34

    9/24/2009

    6

    -

    Wavelength-RoutedNetworks

    Provide lightpaths

    Problems:

    Low bandwidthefficiency

    Large granularity

    Network

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    11

    1

    A

    Wavelength-routing switch

    Access (client) node (e.g., IP router):

    contains (tunable) transmitters and receivers

    Optical circuit switching

    Solvin LTD and RWA roblems

    A lightpath corresponds to a circuit

    Set-up a lightpath

    The whole lightpath is available during the connection Disconnect

    Network elements

    Fiber

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    12

    Optical line amplifier (OLA)

    Optical line terminal (OLT)

    Optical add-drop multiplexer (OADM)

    Optical cross-connect (OXC)

  • 8/2/2019 24 Lecture Wosinska

    7/34

    9/24/2009

    7

    Definitions

    Logical (virtual) topology

    Physical topology: Fiber topology

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 13

    Kind of time multiplexing: packing a low speed

    channels into higher speed channels.

    ObjectiveSeattle

    New York

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 14

    Given a traffic matrix (a forecast) and a fiber (physical) topology:design the networkthat fits the traffic forecast or/andoptimize the (existing) network

  • 8/2/2019 24 Lecture Wosinska

    8/34

    9/24/2009

    8

    Network design

    the fiber (physical) topology and

    traffic matrix (obtained by forecasting): in packets/second

    Determine

    the lightpath (virtual or logical) topology (Lightpath topologydesign: LTD): grooming

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    15

    physical routes through the network and wavelengthassignment (RWA): map the LTD into the physical topology

    Heuristic solution

    ar to eterm ne t e g tpat topo ogyjointly with the routing and wavelength

    assignment Split into separate LTD and RWA problems

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 16

    obtained LTD within the optical layer (i.e. for theobtained LTD solve RWA problem).

  • 8/2/2019 24 Lecture Wosinska

    9/34

    9/24/2009

    9

    LTD:Lightpath topology design

    ven: The traffic demands

    The maximum number of ports per client node

    Lightpaths interconnect client nodes bidirectionally

    Determine the topology and routing of packets

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    17

    Objective (an optimization problem): Minimize the maximum load that any lightpath must

    carry

    RWA:Routing and wavelength assignment

    to the fiber topology

    Objective

    Offline: for all lightpaths determined by the

    LTD minimize the number of wavelen ths

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    18

    used per link

    Online: for demands coming during

    operation minimize the blocking probability

  • 8/2/2019 24 Lecture Wosinska

    10/34

    9/24/2009

    10

    RWA approaches

    Heuristic Routing sub-problem

    Wavelength assignment sub-problem

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 19

    Routing sub-problem

    Fixed shortest-path routing

    Fixed-alternate routing Routing table at each node that contains a list of

    fixed routes to each destination

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    20

    The route is chosen dynamically, depending of the

    network state

    Fault-tolerant routing

  • 8/2/2019 24 Lecture Wosinska

    11/34

    9/24/2009

    11

    Routing algorithms

    Shortest Path selects the shortest source-destinationpath (# of links/nodes)

    Least Loaded Routing avoids the busiest links

    Least Loaded Node avoids the busiest nodes

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 21

    WA sub-problem

    -

    Graph coloring

    Dynamic WA sub-problem Random (R) Wavelength Assignment

    First-Fit (FF)

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    22

    Least-Use LU SPREAD

    Max-Used (MU)/PACK

    Least Loaded (LL)

    etc.

  • 8/2/2019 24 Lecture Wosinska

    12/34

    9/24/2009

    12

    Static (offline) WA:Graph coloring

    NC Wavelen th continuit constraint

    Given: all connections and routes

    Objective: assign wavelengths (colors) to each lightpath so as tominimize the number of wavelengths used under the wavelengthcontinuity constraint

    Constr uct a graph G so that each lightpath in the system isrepresented by a node. Undirected edge between nodes in the

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    23

    G if the corresponding lightpaths share a physical link Color t he nods of G such that no two adjacent nodes have the

    same color.

    Dynamic (online) WAalgorithms

    Random (R)

    Pick one with uniform probability

    First -Fit (FF)

    All wavelengths are numbered Assign the first available wavelength

    Least -Used (LU)/ SPREAD

    Select the wavelength that is the least used in the network

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 24

    Max-Used ( MU)/ PACKED

    Select the most used wavelength in the network

    Least- Loaded (LL) designed for multi-fiber networks

    Select the wavelength with largest residual capacity on the most loadedlink along the connection.

  • 8/2/2019 24 Lecture Wosinska

    13/34

    9/24/2009

    13

    WDM network elementsWDM network example

    Optical line terminals

    (OLTs)

    Optical add-drop

    multiplexers (OADMs)

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    25

    p ca cross-connec s(OXCs)

    Optical line amplifiers

    OLT: Optical line terminal

    Trans onders Adaptation from/to access

    links Wavelength and fiber

    Overhead and FEC BER monitoring

    Main cost of OLT

    Multiplexer To mer e the incomin

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    26

    channels

    Amplifier (optional)

    Optical supervisory channel(OSC) Added and terminated

  • 8/2/2019 24 Lecture Wosinska

    14/34

    9/24/2009

    14

    OADMOptical add/drop multiplexer

    To and from equipment at local node

    Remaining channels pass transparently

    Channel selecti on

    Any channel or only some

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    27

    Reconfigurable: software configurable remotely

    One, a few or any number of channels Modularity

    Loss dependence on number of dropped channels

    OXCOptical Cross-connect

    From input to output ports

    From input to output wavelengths

    Does not include input/output OLTs

    Functions

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    28

    Protection switching (rerouting)

    Performance monitoring

    Wavelength conversion

  • 8/2/2019 24 Lecture Wosinska

    15/34

    9/24/2009

    15

    Types of OXCs

    Transparency

    Cost

    Size

    Transparent or opaquecore

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    29

    Electronic signal

    Signal monitoring

    Optical signal

    Transparency

    All-optical OXCs

    Grooming

    Higher demand for lightpaths No aggregation of low bitrate demands

    Wavelength conversion

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    30

    g er oc ng o g pa eman s

    Signal regeneration More constrained routing of lightpaths

  • 8/2/2019 24 Lecture Wosinska

    16/34

    9/24/2009

    16

    All-optical OXCEx. 1: Clos architecture

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    31

    Three stage strict internal non-blocking Clos architecture.

    Size: 128x128

    L. Wosinska, L. Thylen, and R.P. Holmstrom: Large Capacity Strictly Non-Blocking OXCs Based on MEOMS

    Switch Matrices. Reliability Performance analysis, IEEE/OSA JLT, Vol.19, No.8, Aug. 2001

    All-optical OXCEx. 2: WR architecture

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    32

    Strict internal non-blocking wavelength routing architecture.

    Size 128x128

    L. Wosinska L. Thylen, and R.P. Holmstrom: Large Capacity Strictly Non-Blocking OXCs Based on MEOMS

    Switch Matrices. Reliability Performance analysis, IEEE/OSA JLT, Vol.19, No.8, Aug. 2001

  • 8/2/2019 24 Lecture Wosinska

    17/34

    9/24/2009

    17

    All-optical OXCEx. 3: with TWC

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    33J. Chen, A. Jirattigalachote, L. Wosinska and L Thyln, Novel Node Architectures for Wavelength-Routed WDMNetworks with Wavelength Conversion Capability, in Proc. of ECOC08, Brussels, Belgium, September 2008

    Performance evaluation

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    34J. Chen, A. Jirattigalachote, L. Wosinska and L Thyln, Novel Node Architectures for Wavelength-Routed WDMNetworks with Wavelength Conversion Capability, in Proc. of ECOC08, Brussels, Belgium, September 2008

  • 8/2/2019 24 Lecture Wosinska

    18/34

    9/24/2009

    18

    Shortcomings with OCS

    Low utilisation of resources

    Hard optimization problems need to be solved

    (LTD and RWA)

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    35

    Solution: Optical Packet Switching (?)

    Optical packet switching

  • 8/2/2019 24 Lecture Wosinska

    19/34

    9/24/2009

    19

    OPS Networks

    Large capacity

    High bandwidth efficiency

    Rich routing functionalities

    Great flexibility andreliability

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    37

    Optical packet switching

    Advanta es Complements WDM

    Allows grooming in optical domain

    Allows statistical multiplexing

    Can improve bandwidth utilization within the optical layer

    Increase flexibility

    Problems

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    38

    Technological problems

    Optical control functions

    Synchronization

    Optical buffering

    High complexity of OPS nodes high cost and low reliability

  • 8/2/2019 24 Lecture Wosinska

    20/34

    9/24/2009

    20

    Optical router (OPS node):Needed functions.

    Decoding of packet header Could be electronic: header encoded at lower bit rate

    Setup of switch fabric Packet delayed until setup done (a fixed delay)

    Setup requires scheduling of packets from all inputs Simplified for fixed packet size and synchronized operation

    Fast reconfiguration of fabric (200 ns for 250 byte packet at 10 Gb/s)

    Synchronization: Elastic buffering of packets to align packets at all inputs

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    39

    n y nee e w en sw c a r c s sync ronous Synchronous fabric has better throughput

    Multiplexing of lower-speed streams (and reverse operation, i.e.demultiplexing)

    Contention resolution (e.g. buffering of packets if output busy)

    Contention resolution in OPS

    Contention may be dealt with in

    Time

    Wavelength

    Space

    Electronic packet switching typically rely on the time

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    40

    domain by means of queuing

    What about optical packet switching ?

    Queuing in optical domain is difficult

  • 8/2/2019 24 Lecture Wosinska

    21/34

    9/24/2009

    21

    If output busy:Handling packet contention Drop a packet

    r y v r

    Deflect the packet Send it on a free output

    Restrict the deflection Output that leads to destination

    Output with a route to the destination that is at most mhops longer

    Also called hot-potato routing

    Increases delay and network load

    Creates variable delays and potentially reordering

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    41

    C ange t e wave engt TWC Chose a wavelength available at the output

    All-optical tunable wavelength converters not available yet

    Buffer the packet

    Store the packet until the output is available

    Applying TWC may allow for decrease of the buffer size

    Contention resolutiontechniques

    Bufferless architecturesOCIC

    e ec on rou ng

    TWC

    Electronic buffers

    Optical buffers Placement at a node

    Output buffer

    Input buffer

    switchfabric

    OC

    OC

    IC

    IC

    switchfabric

    IC

    IC

    IC

    OC

    OC

    OC

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    42

    ec rcu a on u er

    Dedicated or shared buffers

    Technology

    FDLs

    Novel optical memories

    EIT or Opt. resonators

    buffer

  • 8/2/2019 24 Lecture Wosinska

    22/34

    9/24/2009

    22

    OPS with electrical buffer

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 43

    L. Wosinska and G. Karlsson, A photonic packet switch for high capacity optical networks, in Proc. NFOEC02, Dallas,Texas, September 2002

    COMPARISON

    Architecture I II

    Buffer Dedicated Shared

    Packet loss probability High Low

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 44

    Scheduling Simple Complex

    L. Wosinska and G. Karlsson, A photonic packet switch for high capacity optical networks, in Proc. NFOEC02, Dallas,Texas, September 2002

  • 8/2/2019 24 Lecture Wosinska

    23/34

    9/24/2009

    23

    Comparison. N=20

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 45

    L. Wosinska and G. Karlsson, A photonic packet switch for high capacity optical networks, in Proc. NFOEC02, Dallas,Texas, September 2002

    A2A1

    Optical buffering

    Fiber delay lines (FDLs)

    Not random access

    Require synchronization

    Supported packet format Constant packet size

    Some configurations support variable packet size A certain granularity

    Not compatible with packet formats of different packet size

    Long fiber delay lines Not very practical solution

    Ex.: For packets containing 53 bytes (ATM cell) at 2.5 Gb/s the length of fiber in the FDLs needs

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    46

    to be the multiples of 640 m

    Feed-forward or feed-back configurations

    Novel solutions for optical memory

    Material subjected to EIT (Electromagnetically induced transparency)

    Optical Cavities

  • 8/2/2019 24 Lecture Wosinska

    24/34

    9/24/2009

    24

    1 km

    Fiber delay line

    Cheap and easy to manufacture

    Several kilometers long

    -

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    47

    No flexibility in terms of storage time

    Requires synchronization

    Many architectures proposed to introduce variabledelay

    EITNovel type for optical memory

    artificially created spectral window of transparencyused to slow and spatially compress light pulses.

    Inside the memory cell, light is converted into a spinexcitation of atoms and its velocity drops down tozero once the coupling beam is turned off. After the

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    48

    coherence is converted back into light signal.

  • 8/2/2019 24 Lecture Wosinska

    25/34

    9/24/2009

    25

    EIT, cont.

    beam

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    49

    Memory cellArriving IP packet

    Optical fiber

    EIT, cont.

    Phase 1 : w r i t i ng

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    50

    Light slows down inside the cell andis spatially compressed

  • 8/2/2019 24 Lecture Wosinska

    26/34

    9/24/2009

    26

    EIT, cont.

    Phase 1 : w r i t i ng

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    51

    Cell length

    The memory cell needs to be long enough to fit the entire packet

    IP packet of 1500bytes at 2,5Gb/s is 1,4 km long in free space andabout 1km long in an optical fiber

    EIT, cont.

    beam is turnedoff

    Phase 2 : s tor age

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    52

    The light is stored in the material

  • 8/2/2019 24 Lecture Wosinska

    27/34

    9/24/2009

    27

    EIT, cont.

    beam is turnedback on

    Phase 3 : read in g

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    53

    The light is recoveredand leaves the cell

    EIT, cont.

    Variable couplingpower

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    54

    No storage of light

    We regulate the slowdown factor by varying the coupling power

    The packet is slowed down in the cell

  • 8/2/2019 24 Lecture Wosinska

    28/34

    9/24/2009

    28

    EIT, cont.

    material slow dow n factor storage t ime

    Quantum dots 40 in room temperature

    107 in very low temperature

    8.7ns

    Atomic vapor 105 up to 0.5 msdepends on thegas

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    55

    Slow down factor and storage time depend on thematerial, temperature, coupling power, bandwidthand wavelength

    Optical cavities use optical resonance in photonic structures

    Optical cavities

    Slow down factor of 104 (depending on the number ofside cavities)

    Storage time: 50 ns

    Chip scale implementation of the system foreseeable

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    56

  • 8/2/2019 24 Lecture Wosinska

    29/34

    9/24/2009

    29

    Requirementsfor optical memory

    Technology Compression rate (cell size) Tuning of the intensity of thecontrol field Temperature and mechanical stress Cost

    Te lecommun ica t ions Wavelength Attenuation anddistortion Bandwidth Packet length Control memory cellsseparately

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    57

    Qo S The storage time Pulse distortion Priority classes

    Comparison

    storage t ime cell size temperat ure bandw idt h-wavelength

    EIT

    Up to 0.5 ms Order of cm Close to 0K or80C

    Depends on thematerial

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    58

    Opticalcavities

    Order of ns Size of a chip Room temp. No limitations

  • 8/2/2019 24 Lecture Wosinska

    30/34

    9/24/2009

    30

    Parallel electricaland optical buffer

    Example 1:OPS with hybrid buffer.

    positions

    Opticaloutputs

    Opticalinputs

    Opticaldemultiplexer

    Switchingmatrix

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    59L. Wosinska, J. Haralson, L. Thyln, Benefit of Implementing Novel Optical Buffers

    in an Asynchronous Photonic Packet Switch, in Porc. ECOC04, Stockholm, Sept. 2004

    ASSUMPTIONS The traffic load is uniformly distributed between the outputs.

    The traffic load at all inputs is identical.

    .

    .

    .

    1

    N

    1

    B u f f e r

    1

    .

    .

    .

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 60

    The transparency class (i.e. packets that can not beconverted to the electrical signal) represents 20% of the totaltraffic

    N + N

    .

    .

    .

    .

    .

    .

    N

    N + 1

    N

    L. Wosinska, J. Haralson, L. Thyln, Benefit of Implementing Novel Optical Buffers

    in an Asynchronous Photonic Packet Switch, in Porc. ECOC04, Stockholm, Sept. 2004

  • 8/2/2019 24 Lecture Wosinska

    31/34

    9/24/2009

    31

    Without output buffer,N > 16

    Packet loss probability

    0 3 5

    0,37

    0,39

    0,41

    0,43

    0,45

    ,

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school 61

    0,31

    0,33

    ,

    0,4 0,5 0,6 0,7 0,8 0,9 1

    load

    simulated, const. packet l ength simulated. IP-traffic calculated

    L. Wosinska, J. Haralson, L. Thyln, Benefit of Implementing Novel Optical Buffers

    in an Asynchronous Photonic Packet Switch, in Porc. ECOC04, Stockholm, Sept. 2004

    IP traffic trans arenc class

    Simulation results.IP traffic.

    1,E-02

    1,E-01

    1,E+00

    packetlossp

    r

    obability

    load 0.2

    load 0.5

    load 0.7

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    62

    The loss probability goes down to a certain point and thanstays constant as the buffer increases.

    1,E-03

    0 2 4 6 8 10 12 14 16

    number of optical buffer positions

    L. Wosinska, J. Haralson, L. Thyln, Benefit of Implementing Novel Optical Buffers

    in an Asynchronous Photonic Packet Switch, in Porc. ECOC04, Stockholm, Sept. 2004

  • 8/2/2019 24 Lecture Wosinska

    32/34

    9/24/2009

    32

    ATM traffic, transparency class

    Simulation results.ATM traffic.

    1,E-06

    1,E-05

    1,E-04

    1,E-03

    1,E-02

    1,E-01

    1,E+00

    loss

    probability

    3 buffer positions5 buffer positions10 buffer positions

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    63

    The lowest achievable packet loss probability for a givennumber of buffer positions reaches a limit that cannot beovercome by increasing the maximum storage time.

    0 0,05 0,1 0,15 0,2 0,25

    maximum storage time in s

    L. Wosinska, J. Haralson, L. Thyln, Benefit of Implementing Novel Optical Buffers

    in an Asynchronous Photonic Packet Switch, in Porc. ECOC04, Stockholm, Sept. 2004

    ATM traffic, trasnparency class

    Simulation results.ATM traffic.

    1,E-06

    1,E-05

    1,E-04

    1,E-031,E-02

    1,E-01

    1,E+00

    packetloss

    pro

    bability load 0.2

    load 0.5

    load 0.7

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    64

    Storage time of 0.5ms is enough to obtain any value of lossprobability for any traffic load.

    0 2 4 6 8 10 12

    number of optical buffer positions

    L. Wosinska, J. Haralson, L. Thyln, Benefit of Implementing Novel Optical Buffers

    in an Asynchronous Photonic Packet Switch, in Porc. ECOC04, Stockholm, Sept. 2004

  • 8/2/2019 24 Lecture Wosinska

    33/34

    9/24/2009

    33

    Example 2:OPS with TWCs and buffer

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    65JiaJia Chen and L. Wosinska, Novel Architectures of Asynchronous Optical Packet Switch, in Proc. of EuropeanConference on Optical Communication ECOC07, Berlin, Germany, September 2007

    Evaluation

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    66JiaJia Chen and L. Wosinska, Novel Architectures of Asynchronous Optical Packet Switch, in Proc. of EuropeanConference on Optical Communication ECOC07, Berlin, Germany, September 2007

  • 8/2/2019 24 Lecture Wosinska

    34/34

    9/24/2009

    Summary Switched networks

    Photonic circuit switching WDM network design: solving offline LTD and RWA

    Objective: minimize number of wavelengths

    Dynamic scenario

    Oblective: minimize blocking probability

    WDM network elements

    Photonic packet switching

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    67

    Makes networks more efficient Many technical challenges for switch design

    Buffering for contention resolution

    Scheduling for contention resolution with possible deflection

    Switching speeds

    Krakow Sept. 29, 2009 L. WosinskaTutorial at BONE Summer school

    68


Recommended