An Evaluation of Scalable Application-level Multicast Built Using - - PowerPoint PPT Presentation

▶
an evaluation of scalable application level multicast
SMART_READER_LITE
LIVE PREVIEW

An Evaluation of Scalable Application-level Multicast Built Using - - PowerPoint PPT Presentation

Apr 08, 2024 131 likes •318 views

Data Centric Networking (R202) paper An Evaluation of Scalable Application-level Multicast Built Using Peer-to-peer Overlays authors: M. Castro et. al. MPhil in ACS reviewer/presenter: S. Trajanovski ( st508 )

slide-1
SLIDE 1

paper

An Evaluation of Scalable Application-level Multicast Built Using Peer-to-peer Overlays

authors: M. Castro et. al.

MPhil in ACS reviewer/presenter: S. Trajanovski (st508)

Data Centric Networking (R202)

slide-2
SLIDE 2

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

2

  • Lack of IP multicast deployment
  • Difference in used overlay
  • Tree building
  • Flooding
  • Difference in routing methods
  • Generalized hypercube routing (Pastry)
  • Cartesian hyper space (CAN)
  • Paper structure..

Motivation Overview of p2p multicast classification

slide-3
SLIDE 3

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

3

  • Topology organization
  • d – dimensional cube
  • Each node owns “its space”
  • Node’s addition
  • Selected region is split into 2 parts
  • Tunable parameters
  • Node dimension, multiple node/zone
  • Net. Aware routing, uniform partitioning etc.

Different routing methods (CAN) P2P overlay networks

slide-4
SLIDE 4

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

4

(before) (after) 2-D CAN topology & node addition

P2P overlay networks Different routing methods (CAN)

slide-5
SLIDE 5

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

5

  • Pastry namespace (128-bits)
  • nodeId & (message, dest_key)
  • Next node -> closest to the dest_key
  • Routing principle
  • 128/b levels and 2 entries each
  • Next node – sharing max. bits with destination
  • At least b bits “closer to destination”

Different routing methods (Pastry) P2P overlay networks

b

slide-6
SLIDE 6

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

6

Routing (65a1fc -> d467c4)

P2P overlay networks Different routing methods (Pastry)

Routing table

slide-7
SLIDE 7

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

7

  • Flooding
  • Main concept: broadcasting
  • Groups: smaller subsets
  • CAN flooding: “naive” & CAN Multicast
  • Pastry flooding: level tagging and forwarding
  • Tree – based (Scribe approach)
  • Firstly target group roots
  • Decentralized approach is scalable

Overlay based application level multicast P2P overlay networks

slide-8
SLIDE 8

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

8

tree-build architectures Different overlays approaches flooding P2P overlay networks

slide-9
SLIDE 9

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

9

  • Experimental environment
  • measured -> number of packets
  • 5 network topologies
  • 1st experimental set: one multicast group
  • 2nd experimental set: multiple multicast groups
  • Experimental phases
  • groups subscription
  • message is multicast to each group

Evaluation Experimental setup

slide-10
SLIDE 10

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

10

  • Experimental criteria
  • Relative (ratio app. level/IP multicast values)
  • Relative delay penalty
  • RMD (maximum ratio), RAD (average ratio)
  • Link stress
  • Number of packets over the link
  • Node stress
  • Routing table size (# nodes) & messages received
  • Message duplication

Evaluation Experimental setup

slide-11
SLIDE 11

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

11

CAN Results CAN flooding CAN tree-build

  • Parameters
  • Number of dimensions (d); nodes per zone (z)
  • policy: (distance, ratio, NDR), uniform part. (on/off)
slide-12
SLIDE 12

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

12

Pastry Results Pastry flooding Pastry tree-build

  • Parameters
  • b - # number of “matched” dest. bits (b= { 1,2,3,4} )
  • TART & TOP
slide-13
SLIDE 13

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

13

More results flooding in CAN

  • Link Stress

CAN tree-based Pastry flooding for b= 4

slide-14
SLIDE 14

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

14

Comparative Results

RMD with localized members

for multiple multicast groups (CDF functions)

RMD with distributed members RMD with both localized and distributed members

slide-15
SLIDE 15

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

15

  • Tapestry and Chord
  • Similar approaches
  • Bayeux and Overcast
  • Different concepts
  • Prospective future work
  • Non scalable
  • End System Multicast, RONs, ISIS
  • IP Multicast

Related Work

slide-16
SLIDE 16

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

16

  • First head2head p2p analysis (4 comb.)
  • Flooding & tree building
  • Hypercube & Cartesian metric space
  • Tree-based is better than flooding
  • Multicast trees-build
  • Pastry better than CAN
  • Flooding overlay costs more..
  • Related work (further considerations)
  • Overcast, Bayeux, IP Multicast

Summary

slide-17
SLIDE 17

Data Centric Networking (R202) student: Stojan Trajanovski (st508)

17

  • Questions??
  • Discussion..