We propose a model for performance evaluation (average throughput, packet delay and loss probability) of IP networks that are dominated by congestion-controlled (i.e., TCP-like) traffic. Our model includes heterogeneous TCP flows, UDP flows, short-lived TCP flows, and TCP flows in a differentiated services network. We also introduce a practical approach to estimating the performance of large sc...

Traffic marking is an essential component in any Differentiated Services (DS) architecture. Traffic marking can be done with respect to a single flow or a flow aggregate. In a DS network, traffic marking is typically performed at aggregate levels, rather than on the level of individual flows. The use of aggregation raises a number of concerns, such as fairness concern. When we deal with aggrega...

In this paper, we propose to adapt the Early Regulation of Unresponsive Flows (ERUF) to third generation wireless networks employing link layer retransmissions. Wireless channel degradations may result in backlog of downlink packets at the link layer queue, causing real-time packets with hard delivery deadlines to expire and be dropped at the receiver. We propose to regulate these congested flo...

In this paper, we analyze the stability of XCP (eXplicit Control Protocol) in a network with heterogeneous XCP flows (i.e., XCP flows with different propagation delays). Specifically, we model a network with heterogeneous XCP flows using fluid-flow approximation. We then derive the conditions that XCP control parameters should satisfy for stable XCP operation. Furthermore, through several numer...

10 The global network of subsea fibre-optic cables plays a critical role in the world economy and is 11 considered as strategic infrastructure for many nations. Sediment density flows have caused 12 significant disruption to this network in the recent past. These cable breaks represent the only 13 means to actively monitor such flows over large oceanic regions. Here, we use a global cable break...

The modern industrial network traffic consists of an increasing amount of soft real-time flows, which can tolerate very small delay and losses. Allowing probabilistic Quality of Service (QoS) violation for these flows can greatly help to improve resource utilization. However, the improvement depends on the statistical properties of competing independent flows. The notion of effective bandwidth ...

We design an efficient on-line approach, FlowMate, for partitioning flows at a busy server into flow groups that share bottlenecks. These groups are periodically input to congestion coordination, aggregation, load balancing, admission control, or pricing modules. FlowMate uses inband packet delay measurements to the receivers to determine shared bottlenecks among flows. Packet delay information...

Workload patterns in Data Center Networks (DCNs) vary by time. As such, deployment of reconfigurable topologies to handle these varying workloads is a promising approach. Network efficiency can be improved by tailoring the network topology to the workload, which can reduce the average path length thus supporting more flows simultaneously. However, without proper software support, physically alt...

Application performance and availability can be improved by aggressive background replication – distributing data across a network to where it is needed before it is requested. However, given the rapid fluctuations of available network bandwidth and changing resource costs due to technology trends, hand tuning applications risks (1) complicating applications, (2) being too aggressive and interf...