ITC-21 IGen: Generation of Router-level Internet Topologies Using - - PowerPoint PPT Presentation

itc 21
SMART_READER_LITE
LIVE PREVIEW

ITC-21 IGen: Generation of Router-level Internet Topologies Using - - PowerPoint PPT Presentation

ITC-21 IGen: Generation of Router-level Internet Topologies Using Network Design Heuristics B. Quoitin , V. Van den Schrieck, P. Franois and O. Bonaventure (bruno.quoitin@uclouvain.be) 1 (C) 2009, B. Quoitin, UCLouvain, Belgium 1 ITC-21,


slide-1
SLIDE 1

1

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

1

ITC-21

IGen: Generation of Router-level Internet Topologies Using Network Design Heuristics

  • B. Quoitin, V. Van den Schrieck, P. François and O. Bonaventure

(bruno.quoitin@uclouvain.be)

slide-2
SLIDE 2

2

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

2

Motivations

  • Routing Sensitivity Analysis
  • Study sensitivity of (BGP) routing to several

“topological” parameters

Topology Routing Solver (C-BGP) Selected Paths Network Design Parameters Topology Generator Routing Efficiency Metrics Measure Routing Efficiency Relate Sensitivity of Metrics to Parameters ???

http://inl.info.ucl.ac.be/softwares/c-bgp

slide-3
SLIDE 3

3

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

3

Generating Topologies...

  • Issue 1
  • “[...] a careful design process [...] can yield high

performance topologies, but these are extremely rare from a probabilistic graph point of view.”

[Lun Li et al, ACM SIGCOMM'04]

Set of all Possible Topologies Subset of “Designed” Topologies Probabilistic models sample this set Likelihood Performance

slide-4
SLIDE 4

4

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

4

Generating Topologies...

  • Issue 2
  • Well-known topology generator: Waxman
  • How to set

to achieve realistic network design

  • bjectives ?

– minimize delay – maximize bandwidth – ensure robustness – limit financial cost – match equipment / technology constraints – ...

Pu ,v=.e

−d u ,v . L

 ,

slide-5
SLIDE 5

5

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

5

Methodology (1)

Input: Set of nodes Group into PoPs Generate backbone and PoPs Assign capacities and weights Build iBGP topology

  • Import existing set of nodes
  • with coordinates (e.g. from existing network or geoloc DB)
  • Generate set of nodes
  • (X,Y): X~Uniform, Y~Uniform
  • optional constraints: (x,y) falls into set of polygons
slide-6
SLIDE 6

6

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

6

Methodology (2)

Input: Set of nodes Group into PoPs Generate backbone and PoPs Assign capacities and weights Build iBGP topology

  • Clustering methods
  • based on distance (euclidian or geodesic)
  • k-medoid (targets k PoPs) or hierarchical (target k PoPs but

bounds intra-PoP variance)

slide-7
SLIDE 7

7

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

7

Methodology (3)

Input: Set of nodes Group into PoPs Generate backbone and PoPs Assign capacities and weights Build iBGP topology

  • Mesh synthesis methods
  • trees, union of disjoint trees, rings, union of rings,

triangulation

  • loose control on graph density and connectedness
  • specific layout for PoPs (e.g. Sprint-like)
slide-8
SLIDE 8

8

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

8

Backbone Generation Methods

MST Hybrid MST / SPT MENTOR (0.3) Hybrid MST / SPT MENTOR (0.7) Union of 2 disjoint trees (MSTs in example) Delaunay triangulation Minimum cost cycle Union of 2 rings Union of 3 rings

There are 2 links here :-)

slide-9
SLIDE 9

9

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

9

PoP Layout (Sprint-like)

1st param: number of backbone routers in PoP (default: 2) 2nd param: number of links from access routers to backbone routers

slide-10
SLIDE 10

10

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

10

Methodology (4)

Input: Set of nodes Group into PoPs Generate backbone and PoPs Assign capacities and weights Build iBGP topology

  • Capacities
  • Based on traffic matrix (computes APSP and forward)
  • 2-levels. Example core=1GB / access=155MB
  • IGP weights
  • Based on distance / inverse of capacity
slide-11
SLIDE 11

11

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

11

Methodology (5)

Input: Set of nodes Group into PoPs Generate backbone and PoPs Assign capacities and weights Build iBGP topology

  • iBGP topology
  • full-mesh
  • iBGP hierarchy (2 levels), route-reflectors chosen among

backbone routers, access routers are clients of route- reflectors in their PoP

slide-12
SLIDE 12

12

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

12

Example (1)

http://inl.info.ucl.ac.be/softwares/igen

Generate 50 vertices Constrain /w polygon

slide-13
SLIDE 13

13

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

13

Example (2)

1). Build 5 clusters based on distance 3). Build backbone using triangulation 2). Build PoPs using “Sprint” layout 4). Assign IGP weights based on link mileage 5). Assign 155M/1G to access/backbone links resp. k-medoids:5 sprint:2:2 delaunay distance:0 access-backbo

slide-14
SLIDE 14

14

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

14

Example (3)

One of the 5 PoPs (clusters)

slide-15
SLIDE 15

15

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

15

Conclusion

  • Further work
  • Finer control on graph density
  • More than 2-tiers
  • Other typical PoP designs
  • Other node placement strategies
  • Synthesis of Internet-wide topologies (in progress)

Set of all Possible Topologies Subset of “Designed” Topologies Our method generates a few elements. How to better sample subset ?

slide-16
SLIDE 16

16

(C) 2009, B. Quoitin, UCLouvain, Belgium ITC-21, Paris, September 2009

16

Questions

???