Virtual Topology Adaptation in WDM Mesh Networks

1
Virtual Topology

• Wavelength routed network
• Lightpath a direct connection between nodes.
• Physical topology ? Virtual topology

2
VT Design

• Given
• Physical topology G(V,E)
• Traffic intensities between nodes.
• Find
• Lightpaths to be set-up.? Virtual topology
• Objective
• Maximize traffic load.
• Minimize cost.

3
VT Reconfiguration

• Traffic is dynamic Traffic intensities change
  in time.? VT may be inefficient for new
  traffic.
• Problem How can we change the VT to accommodate
  the new traffic intensities?
• Solutions
• Redesign the VT.
• Change some lightpaths ? Reconfiguration.

4
Illustrative Example
0 1
0 1 2 3 0 0 1 0 0 1 0 0
1 0 2 0 0 0 1 3 1 0 0 0
3 2
0 1 2 3 0 0 0 0 1 1 1 0
0 0 2 0 1 0 0 3 0 0 1 0
5
Motivation
t1 t2

• Backbone network
• Slow changes in traffic ? Change the virtual
  topology slowly.

6
Earlier Approaches

• Wait until VT becomes inefficient.
• Run an optimization algorithm ? new VT.
• Run a transformation algorithm.Old VT
  New VT
• Branch-exchange method

7
Our Approach

• Observe the traffic Adapt VT
• New VT adjacent to old VT Only 1 lightpath is
  different.
• Simple reconfiguration.
• Simple transition.

continuously.
8
Optimization Goals

• Using minimum number of lightpaths.
• Load balancing.
• Minimizing the number of changes? transition
  periods length? traffic rerouting

9
Main Idea

• Add or delete 1 lightpath at a time.
• Add a lightpath if some lightpaths are heavily
  loaded.
• Delete a lightpath if some lightpaths are lightly
  loaded.
• 2 parameters
• High watermark
• Low watermark

10
Implementation

• Traffic loads on links are observed
  periodicallyObservation period T.
• T should be small enough to follow the changes.
• Typically few 100s

11
Adaptation Algorithm

• Start with initial topology.
• Every T seconds do
• Compare the load Li on each
• lightpath i with the watermarks.
• if Li gt WH then
• add an appropriate lightpath.
• else if Li lt WL then
• delete an appropriate lightpath.
• else
• do not adapt the VT.

12
MILP

• Objective Minimize Lmax
• Constraints

(1)
13
MILP (cont.)

• Constraints
• if Lmax gt WH then
• if Lmin lt WL then

(2)
14
Comparison

• We compare MILP to a full-reconfiguration method.

Start with initial topology. Every T seconds
do Find the minimum lightpaths for the
new traffic pattern. Find the virtual topology
such that It has minimum lightpaths.
It requires minimum changes from the
previous topology.
15
Results (MILP)
N 6 W 4 Tx 4 WH 0.8
16
Results (MILP)
17
Results (Heuristic)
18
Results (Heuristic)
19
Results (Heuristic)