Brief description
At each time interval, this algorithm adjusts dynamically the number of lightpaths for each demand with respect to the ones in the previous period.
In case that new lightpaths are required, it performs an RWA using first-fit (wavelength continuity is enforced) over a set of precomputed k-shortest paths (maximum path length in km equal to maxLightpathLengthInKm). These paths are evaluated in order. Once a feasible RWA is found, then the lightpath is allocated, and no other candidate solution is checked.
Algorithm description table
| Algorithm inputs | Receives the current traffic demand volume. Algorithm parameters:
|
|---|---|
| Algorithm outputs | The number of lightpaths for each demand is adjusted to match the current traffic requirements |
| Required libraries | None |
| Keywords | WDM |
| Authors | Pablo Pavón Mariño, José Luis Izquierdo Zaragoza |
| Date | March 2014 |
| Code | TVSim_AA_WDM_basicTrafficAllocationAlgorithm.java |
Detailed description
Let \( currentNumLps_d \) be the number of lightpaths which are active for demand \( d\in D \) in the current time, \( h_d \) the new traffic volume in Gbps for demand \( d\in D\), and \( binaryRatePerChannel\_Gbps \) the binary rate per lightpath. Then, the number of lightpaths for demand \( d\in D \), denoted as \( numLps_d \), for the new period is computed as follows:
\( numLps_d = \left \lceil{\frac{h_d}{binaryRatePerChannel\_Gbps }}\right \rceil \)
Then, we adjust the number of lightpaths as follows:
- If \( numLps_d - currentNumLps_d > 0 \), add \( numLps_d - currentNumLps_d \) lightpaths using the pre-computed k-shortest paths and first-fit wavelength assignment
- If \( numLps_d - currentNumLps_d < 0 \), remove the latest \( currentNumLps_d - numLps_d \) lightpaths from demand
- Otherwise, do nothing
