Richa Malhotra

A feedback fluid queue with two congestion control thresholds

Feedback fluid queues play an important role in modeling congestion control mechanisms for packet networks. In this paper we present and analyze a fluid queue with a feedback-based traffic rate adaptation scheme which uses two thresholds. The higher threshold B1 is used to signal the beginning of congestion while the lower threshold B2 signals the end of congestion. These two parameters together allow to make the trade-off between maximizing throughput performance and minimizing delay. The difference between the two thresholds helps to control the amount of feedback signals sent to the traffic source. In our model the input source can behave like either of two Markov fluid processes. The first applies as long as the upper threshold B1 has not been hit from below. As soon as that happens, the traffic source adapts and switches to the second process, until B2 (smaller than B1) is hit from above. We analyze the model by setting up the Kolmogorov forward equations, then solving the corresponding balance equations using a spectral expansion, and finally identifying sufficient constraints to solve for the unknowns in the solution. In particular, our analysis yields expressions for the stationary distribution of the buffer occupancy, the buffer delay distribution, and the throughput.

Optimized routing for providing ethernet LAN services

Ethernets-move into metropolitan and wide area networks is driving a rapidly growing market opportunity. Current Ethernet services come in two basic flavors, Ethernet line and LAN, providing point-to-point and multipoint connectivity, respectively. The LAN services, although more cost effective, are lagging behind in deployments due to associated QoS and bandwidth provisioning issues. The Ethernet service provider needs to provision the network to meet current and future traffic demands, where the traffic is unpredictable and bursty. The goal is to minimize. overprovisioning and complexity. To add to the challenge, Ethernet routing is based on simple self-learning and relies on spanning tree routing. In this article we propose an Ethernet-specific load balanced routing mechanism, which is robust to dynamic traffic demand. It significantly reduces overprovisioning, is simple and static, and requires only bandwidth profiles associated with service level agreements at the ingress and egress links. Our simulation results show that our scheme provides performance improvements over a recently proposed approach for switched Ethernet as well as a related load balancing approach for Ethernet over MPLS networks.

Modeling the interaction of IEEE 802.3x hop-by-hop flow control and TCP end-to-end flow control

Ethernet is rapidly expanding beyond its niche of local area networks. However, its success in larger metropolitan area networks will be determined by its ability to combine simplicity, low costs and quality of service. A key element in successfully transporting bursty traffic and at the same time providing QoS, is congestion control. The Ethernet standard IEEE 802.3x defines a hop-by-hop congestion control mechanism. The performance of this scheme generally depends on its interaction with higher layer application traffic, and especially, with TCP controlled traffic which has its own end-to-end congestion control mechanism. In this paper we focus on the performance modeling and analysis of this interaction. Our model takes into account the influence of various network and traffic parameters. The validity of the proposed model is assessed by comparison of the results to simulations. In our experiments we observe that an increase of the round trip time has a positive influence on the interaction of hop-by-hop and TCP congestion control, and that the use of hop-by-hop flow control is only beneficial when the load is not high.

Managing service-level agreements in metro Ethernet networks using backpressure

Ethernet, which has long been used for local area networks (LANs), is increasingly being employed for metropolitan area networks (MANs). Because it is familiar, simple, and cost-effective, it is becoming popular as a means of transporting data in a MAN. A prerequisite for providing different levels of service in a MAN is the ability to support and maintain service-level agreements (SLAs). Proper management of SLAs ensures that users do not violate their contracts and makes it possible to guarantee quality of service (QoS). Such management requires support for traffic flow control, monitoring, and policing. This paper presents a novel Ethernet traffic-policing mechanism based on backpressure and able to provide feedback to an end user or access network when SLAs are violated, instead of simply dropping packets. The performance of this mechanism is analyzed by means of tests. The tests focus on the effects of the mechanism on throughput and fairness for transmission control protocol (TCP) traffic and on jitter and packet drops for user datagram protocol (UDP) multimedia traffic.

Resource allocation in a multiple-priority buffered link

In this paper we consider a multiple-priority buffered link, with a focus on resource allocation issues; our study was motivated by such issues in Carrier Ethernet, but the applicability of the results is by no means limited to this technology. In our model, the resource is shared by two priority classes. As the high-priority queue — intended for traffic generated by delay-sensitive applications — typically has a small buffer, the low-priority queue can be modelled as a queue with a time-varying service rate; this service rate behaves independently of the input of the low-priority queue. The analysis of the high-priority queue is standard, and we therefore provide an in-depth study of the performance of the low-priority queue. Assuming that all traffic offered to the system is subjected to a leaky-bucket type of policer, one of the approaches that we present borrows elements from the setup of [6], in which the notion of effective bandwidth plays a crucial role.