Performance modelling and traffic characterisation of optical networks
A review is carried out on the traffic characteristics of an optical carrier's OC-192 link, based on the IP packet size distribution, traffic burstiness and self-similarity. The generalised exponential (GE) distribution is employed to model the interarrival times of bursty traffic flows of IP packets whilst self-similar traffic is generated for each wavelength of each source node in the optical network. In the context of networks with optical burst switching (OBS), the dynamic offset control (DOC) allocation protocol is presented, based on the offset values of adapting source-destination pairs, using preferred wavelengths specific to each destination node. Simulation evaluation results are devised and relative comparisons are carried out between the DOC and Just-Enough-Time (JET) protocols. Moreover parallel generators of optical bursts are implemented and simulated using the Graphics Processing Unit (GPU) and the Compute Unified Device Architecture (CUDA) and favourable comparisons are made against simulations run on general-purpose CPUs.
Item Type | Other |
---|---|
Uncontrolled Keywords | bursty traffic; Compute Unified Device Architecture (CUDA); Dense-wavelength Division Multiplexing (DWDM); Generalised Exponential (GE) Distribution; Generalised Exponential Distribution (GE); Graphics Processing Unit (GPU); Just Enough Time (JET) protocol; Long-Range Dependence (LRD); Optical Burst Switching (OBS); optical burst switching (OBS) protocol; Optical Packet Switching (OPS); Parallel Processing; self-similar traffic; Self-Similarity; Synchronous Optical Networking (SONET); Wavelength division multiplexing (WDM); Wavelength Division Multiplexing (WDM) |
Subjects |
Computer Science(all) Mathematics(all) > Theoretical Computer Science |
Date Deposited | 26 Jul 2024 13:25 |
Last Modified | 26 Jul 2024 13:25 |
Explore Further
Read more research from the creator(s):
Find work associated with the faculties and division(s):
- Centre for Computer Science and Informatics Research
- Department of Computer Science
- School of Physics, Engineering & Computer Science
Find other related resources: