(It's better to create than destroy what's unnecessary)

Wednesday, June 27, 2007

Problems with Chandra's Channel Assignment Paper

  1. It defines a centralized scheme to do vertex coloring
  2. The scheme takes (|V|) ^2 complexity
  3. They have a distributed version which takes (rD|V|^2) complexity. Where 'r' is the radius and D is the degree of neighborhood.
  4. Assumes Orthogonal channels.
  5. For non-orthogonal channels: Paper uses data from a firmware paper where the spectrum of overlapped channels is used to calculate the interference. This does not take into account the overlapped band's usage that actually creates interference.

No comments:

Labels