A distributed congestion avoidance scheme, the End-to-End Rate Control Algorithm (EERCA), has been proposed [5] for Available Bit Rate (ABR) service. In this work, we enhance EERCA by replacing the Queue Occupancy Reduction with the queuing delay control (QDC), which manipulates the queue occupancy more efficiently with less functional complexity. Furthermore, it alleviates per-VC accounting re...

We consider graphs possibly with loops but with no parallel edges. A graph without loops is simple. Let G be a graph with vertex set V (G) and edge set E(G). If S ⊆ E(G) then G[S] denotes the spanning subgraph of G with edge set S. We say G is ordered if V (G) = {1, 2, . . . , |V (G)|}. Let G be an ordered graph. Let `(e) and r(e) denote the endpoints of each edge e ∈ E(G) such that `(e) ≤ r(e)...

The heap-on-top (hot) priority queue data structure 6] improves on the best known times for Dijkstra's shortest path algorithm. It also has very good practical performance and is robust over a wide range of graph types. The heart of Dijkstra's algorithm is a monotone priority queue, that is, a priority queue where no element on the queue ever becomes smaller than the most recently extracted ele...

We study a network of fluid queues in which exogenous arrivals are modulated by a continuous-timeMarkov chain. Service rates in each queue are proportional to the queue size, and the network is assumed to be irreducible. Thequeue levels satisfy a linear, vectorvalued differential equation. We obtain joint moments of the queue sizes recursively, and deduce the Laplace transform of the queue size...

This paper identifies various parametrized dynamic priority queue disciplines from literature which are complete in two class M/G/1 queue. Equivalence between these queue disciplines is established by obtaining non linear transformation in closed form. Motivation behind these completeness and equivalence results is discussed from optimal control perspective. Some comments on global FCFS policy ...

We consider a closed queueing network, consisting of two FCFS single server queues in series: a queue with general service times and a queue with exponential service times. A fixed number N of customers cycles through this network. We determine the joint sojourn time distribution of a tagged customer in, first, the general queue and, then, the exponential queue. Subsequently, we indicate how th...

In this work, we study the scalability, performance, design and implementation of basic data structure abstractions, such as a queue, for next generation multicore systems. We propose two algorithms for concurrent queue. Our first algorithm, a wait-free queue, provides an efficient replacement to a lock-free queue. Lock-free queue is considered very efficient, but does not provide local progres...

Production parallel systems are space-shared and hence employ batch queues in which the jobs submitted to the systems are made to wait before execution. Thus, jobs submitted to parallel batch systems incur queue waiting times in addition to the execution times. Prediction of these queue waiting times is important to provide overall estimates to the users and can also help metaschedulers make sc...