IP全光融合網路

Abstract

With emerging trends for ubiquitous data services, the IP-optical convergence networks are gradually becoming a preferred technology. In optical access networks, Ethernet Passive Optical Network (EPON) is becoming a major technology for pervasive access. For optimizing the performance, Chapter 2 firstly applies IEEE 802.3 frame burst mode (FBM) to the EPON downstream link and compare its performance with current techniques for various traffic patterns. Then, Chapter 3 investigates a preemptive priority transmission technique that works on MAC and PHY layer of EPONs. It effectively reduces the frame queuing delay as well as jitter, so as to enhance precise time synchronization in EPONs. Thus, the hard realtime QoS-guarantee for highly delay-sensitive services such as Circuit Emulation Services over EPON (CESoEPON) could be enhanced. In optical core networks, a major obstacle in realizing all-optical switched IP network is due to the large processing delays of optical switching nodes which still need OEO conversion at the optical header (label) processing. In Chapter 4, a novel concept: SOLOS architecture was proposed to support AOS in word and deed. Similar with GMPLS, an optical label switching based on source routing information is attached to IP packets for fast interconnection purpose. The proposed labeling scheme requires neither table lookup nor label replacing in the stateless all-optical switching node, labeled packets/bursts will be forwarded through self-routing, and the only operation in optical switching nodes is to strip off its label part which represents the output port. Thus the whole processing has been completed in the optical domain. Finally, an innovative all-optical multicasting architecture: COXcast based on unicast and eXplicit multicast (Xcast) was proposed in Chapter 5, the scheme represents the multicast tree by decoding the bitmap of its corresponding output port using a common identifier and a node-specific key, so that multicast packets can be self-routed to multiple receivers without requiring header modification, and intermediate nodes on the multicast tree will be stateless. COXcast can reduce the processing cost, protocol overhead and delivery latency. The simulation results indicated that our proposed scheme achieves a remarkable performance in scalability while simplifying the deployment and management of a large number of medium-scale multicast groups especially for multiparty conferences in AOS networks. The remarkable performance improvement of the proposed schemes and architectures are proven through both mathematical analysis and simulations. The purposes of this dissertation are to shed light on a very active field of research on next generation all-IP photonics telecommunication and Internet, as well as to make contribution on everyone’s modern style of living.

With emerging trends for ubiquitous data services, the IP-optical convergence networks are gradually becoming a preferred technology. In optical access networks, Ethernet Passive Optical Network (EPON) is becoming a major technology for pervasive access. For optimizing the performance, Chapter 2 firstly applies IEEE 802.3 frame burst mode (FBM) to the EPON downstream link and compare its performance with current techniques for various traffic patterns. Then, Chapter 3 investigates a preemptive priority transmission technique that works on MAC and PHY layer of EPONs. It effectively reduces the frame queuing delay as well as jitter, so as to enhance precise time synchronization in EPONs. Thus, the hard realtime QoS-guarantee for highly delay-sensitive services such as Circuit Emulation Services over EPON (CESoEPON) could be enhanced. In optical core networks, a major obstacle in realizing all-optical switched IP network is due to the large processing delays of optical switching nodes which still need OEO conversion at the optical header (label) processing. In Chapter 4, a novel concept: SOLOS architecture was proposed to support AOS in word and deed. Similar with GMPLS, an optical label switching based on source routing information is attached to IP packets for fast interconnection purpose. The proposed labeling scheme requires neither table lookup nor label replacing in the stateless all-optical switching node, labeled packets/bursts will be forwarded through self-routing, and the only operation in optical switching nodes is to strip off its label part which represents the output port. Thus the whole processing has been completed in the optical domain. Finally, an innovative all-optical multicasting architecture: COXcast based on unicast and eXplicit multicast (Xcast) was proposed in Chapter 5, the scheme represents the multicast tree by decoding the bitmap of its corresponding output port using a common identifier and a node-specific key, so that multicast packets can be self-routed to multiple receivers without requiring header modification, and intermediate nodes on the multicast tree will be stateless. COXcast can reduce the processing cost, protocol overhead and delivery latency. The simulation results indicated that our proposed scheme achieves a remarkable performance in scalability while simplifying the deployment and management of a large number of medium-scale multicast groups especially for multiparty conferences in AOS networks. The remarkable performance improvement of the proposed schemes and architectures are proven through both mathematical analysis and simulations. The purposes of this dissertation are to shed light on a very active field of research on next generation all-IP photonics telecommunication and Internet, as well as to make contribution on everyone’s modern style of living.