Martin Köhn

Comparison of contention resolution strategies in OBS network scenarios

Optical burst switching (OBS) has attracted interest as a transport network architecture for the future optical Internet. As OBS relies on statistical multiplexing efficient contention resolution is a key issue in order to achieve a low burst loss probability. Basically, contentions can be resolved by wavelength conversion, deflection routing and delaying the burst in a fiber delay line or a combination of these schemes. This paper compares the basic and combined contention resolution strategies in two reference core network scenarios with respect to burst loss probability and end-to-end transfer delay. We show that the effectiveness of those contention resolution schemes highly depends on the load offered to the network and the dimensioning of specific nodes and links. For high load, contention resolution schemes applying deflection routing have an end-to-end transfer time increase in the order of 10-60% depending on the scheme.

Ethernet – A Survey on its Fields of Application

During the last decades, Ethernet progressively became the most widely used local area networking (LAN) technology. Apart from LAN installations, Ethernet became also attractive for many other fields of application, ranging from industry to avionics, telecommunication, and multimedia. The expanded application of this technology is mainly due to its significant assets like reduced cost, backward-compatibility, flexibility, and expandability. However, this new trend raises some problems concerning the services of the protocol and the requirements for each application. Therefore, specific adaptations prove essential to integrate this communication technology in each field of application. Our primary objective is to show how Ethernet has been enhanced to comply with the specific requirements of several application fields, particularly in transport, embedded and multimedia contexts. The paper first describes the common Ethernet LAN technology and highlights its main features. It reviews the most important specific Ethernet versions with respect to each application fields requirements. Finally, we compare these different fields of application and we particularly focus on the fundamental concepts and the quality of service capabilities of each proposal.

Modeling and performance evaluation of iSCSI storage area networks over TCP/IP-based MAN and WAN networks

This paper provides a concise modeling and performance evaluation of the iSCSI storage area network (SAN) architecture and protocol. SANs play a key role in business continuity, enterprise-wide storage consolidation and disaster recovery strategies in which storage resources are most often distributed over many distant data center locations. In the future, SAN traffic will be transported over IP-based networks, e.g., enterprise virtual private networks, to benefit from converged networks and save cost. In these scenarios, the impact of end-to-end delay and QoS of broadband networks on SAN performance is critical and has to be well understood by IT departments when deploying IP-storage solutions and network operators when designing transport network services for SAN applications. In this context, we propose models for iSCSI write requests over TCP/IP networks, e.g., as used in asynchronous mirroring applications. In addition to the analysis for individual requests we present - to the best of our knowledge for the first time - the evaluation of an iSCSI session under a realistic request traffic model with and without interleaving. We analyze the throughput and total request write times for different network dimensions, i.e., round-trip times, and QoS levels, processing delays in the iSCSI layer as well as request characteristics

Comparison of SDH/SONET-WDM multi layer networks with static and dynamic optical plane

Multi layer networks are a very attractive solu- tion to cope with the increasing dynamics and capacities in todays core networks. In SDH/SONET multi layer networks, client layer SDH/SONET connections are groomed to wave- length channels and transported using end-to-end lightpaths. Also, intermediate grooming can yield to a more efficient utili- zation of network resources. In contrast to many other IP-over- WDM network architectures, a clear and efficient evolutionary path exists to upgrade todays SDH/SONET networks based on point-to-point WDM links to dynamic SDH/SONET-WDM networks. In this paper, two principal SDH/SONET-WDM multi layer network architectures are investigated covering the dynamics either only in the electrical layer or in both layers, respectively. We present a detailed performance evaluation based on simu- lation studies and compare both architectures for different total network capacity and grooming strategies in order to show benefits and drawbacks for the introduction of dynamics in the optical plane of todays backbone networks. We also sys- tematically analyze the traffic composition which until now has only rarely been investigated.

Evaluation of Packet Delay in OBS Edge Nodes

In OBS networks, packets are aggregated into bursts in the edge nodes and then send through the all-optical network to the destination node. Most of the delay phenomenon (variable delay) occurs in the edge node, which is an important issue for the QoS provisioning. The paper characterizes the packet delay in the edge node, which is mainly determined by the burst assembly, transmission queuing as well as the characteristic of the incoming IP traffic

Improving fairness in multi service multi layer networks

Fairness is an important but seldom addressed aspect in multi layer networks with traffic grooming capabilities. Often an excessive high blocking probability is observed for high capacity connections compared to small capacity connections. Similar, connections with more hops are more often blocked than those to the neighbor node. In this paper, we present two extensions of a bandwidth fairness algorithm that is based on a call admission control (CAC). First, this algorithm will be generalized for any class-based system. Second, an improvement is presented in order to reduce the performance drawbacks.

A generic 10 Gbps assembly edge node and testbed for frame switching networks

Packet assembly at the network edge is one solution to reduce the high packet rates in core network switches. For this, specialized edge nodes called Assembly Units are needed that assemble client packets into containers and vice versa. In this paper we present the detailed architecture and implementation of a generic Frame Assembly Unit for the Frame Switching architecture along with the testbed used for validation. Our design supports timer and threshold based assembly including packet fragmentation for fixed and variable size container frames at 10 Gbps per direction. For assembly and packet delineation we use the ITU-T Generic Framing Procedure. We report performance and implementation results for an overall design that operates with a 128 Bit data-path at 100 MHz on Xilinx Virtex4 FPGAs

Optimized Frame Packing for OFDMA Systems

Orthogonal frequency division multiple access (OFDMA) is the basis for several emerging wireless systems, such as 802.16e (WiMAX) or 3GPP long term evolution (LTE). In OFDMA, different users are multiplexed in time and frequency. In the 802.16e adaptive modulation and coding (AMC) downlink, the data bursts for a particular terminal have a rectangular shape and need to be placed in the two-dimensional time/frequency plane. The position and shape of the rectangles is arbitrary, and it is the task of the frame packer to pack the frame efficiently, wasting as little space as possible. In this paper, we treat the frame packing problem as a strip-packing problem. We solve this combinatorial optimization problem by developing a suitable representation for a genetic algorithm. This algorithm can reach within 5% of the theoretical lower bound for the packing efficiency.

Design and Evaluation of a Burst Assembly Unit for Optical Burst Switching on a Network Processor

Optical Burst Switching (OBS) has been proposed in the late 1990s as a novel photonic network architecture directed towards efficient transport of IP traffic. OBS aims at cost-efficient and dynamic provisioning of sub-wavelength granularity by optimally combining electronics and optics. In order to reduce the number of switching decisions in OBS core nodes, traffic is aggregated and assembled to bursts by the Burst Assembly Unit in an OBS ingress edge node. This Burst Assembly Unit is responsible for buffering incoming packets in queues and sending them as bursts as soon as a minimum burst length is reached and/or a timer expires. Typically, dozens of different queues must be able to handle high volumes of traffic.

OBS vs. OpMiGua - A Comparative Performance Evaluation

Optical Burst Switching (OBS) and Optical Migration Capable Networks with Service Guarantees (OpMiGua) are two all-optical network architectures. In this paper we compare both by means of a quantitative performance evaluation based on simulations. In order to achieve a maximum of comparability both models are chosen as similar as possible and especially are fed with identical traffic. Results show differences regarding loss probabilities at which OpMiGua has a better performance.