Arkady Kanevsky

An adaptive, distributed airborne tracking sysem

This paper describes a United States Air Force Advanced Technology Demonstration (ATD) that applied value-based scheduling to produce an adaptive, distributed tracking component appropriate for consideration by the Airborne Warning and Control System (AWACS) program. This tracker was designed to evaluate application-specific Quality of Service (QoS) metrics to quantify its tracking services in a dynamic environment and to derive scheduling parameters directly from these QoS metrics to control tracker behavior. The prototype tracker was implemented on the MK7 operating system, which provided native value-based processor scheduling and a distributed thread programming abstraction. The prototype updates all of the tracked-object records when the system is not overloaded, and gracefully degrades when it is. The prototype has performed extremenly well during demonstrations to AWACS operator and tracking system designers. Quantitative results are presented.This paper describes a United States Air Force Advanced Technology Demonstration (ATD) that applied value-based scheduling to produce an adaptive, distributed tracking component appropriate for consideration by the Airborne Warning and Control System (AWACS) program. This tracker was designed to evaluate application-specific Quality of Service (QoS) metrics to quantify its tracking services in a dynamic environment and to derive scheduling parameters directly from these QoS metrics to control tracker behavior. The prototype tracker was implemented on the MK7 operating system, which provided native value-based processor scheduling and a distributed thread programming abstraction. The prototype updates all of the tracked-object records when the system is not overloaded, and gracefully degrades when it is. The prototype has performed extremenly well during demonstrations to AWACS operator and tracking system designers. Quantitative results are presented.

Improved algorithms for graph four-connectivity

We present a new algorithm based on ear decomposition for testing vertex four-connectivity and for finding all separating triplets in a triconnected graph. The sequential implementation of our algorithm runs in O(n2) time and the parallel implementation runs in O(logn) time using O(n2) processors on a CRCW PRAM, where n is the number of vertices in the graph. This improves previous bounds for the problem for both the sequential and parallel cases. The sequential algorithm is optimal if the input is specified in adjacency matrix form, or if the input graph is dense.

A generalized utilization bound test for fixed-priority real-time scheduling

C.L. Liu and J.W. Layland (1973) provided a utilization bound test which is applicable to rate-monotonic priority assignment, when all tasks deadlines are at the end of their periods. Subsequently, J. Lehoczky et al. (1989) provided an exact schedulability criterion which is applicable to any arbitrary priority assignment scheme, with no restriction on task deadlines, which can be used when all task computation times are known exactly. In this work, we fill an important gap between these two, by presenting a technique for deriving utilization bounds, based on linear programming, that is applicable to similarly generalized situations, but does not require knowledge of exact task computation times.

Embedding rings in recursive networks

The WK-Recursive network (WKRN) is a hierarchical interconnection network that is recursively defined and has excellent properties for scalable message-passing computer systems. In this paper we describe the ability of a WKRN to implement algorithms that use the communication pattern of rings. We first describe how rings of arbitrary size can be embedded in WKRNs. We then describe how Hamiltonian cycles can he embedded in a WKRN in the presence of up to W-3 faulty links. The existing scheme for fault-tolerant embedding of Hamiltonian cycles tolerates up to [(W-3)/2] faulty edges. Thus, the new scheme for embedding Hamiltonian cycles tolerates twice as many faulty links as the existing scheme.<<ETX>>

MPI/RT-an emerging standard for high-performance real-time systems

The past few years saw an emergence of standardization activities for real-time systems including standardization of operating systems (series of POSIX standards), of communication for distributed (POSIX.21) and parallel systems (MPI/RT) and real-time object management (real-time CORBA). The article describes the ongoing work of real-time message passing interface (MPI/RT) standardization. MPI/RT advances the Message Passing Interface Standard (MPI), emphasizing changes that enable and support real-time communication, and is targeted for embedded, fault-tolerant and other real-time systems.

Fixed-priority scheduling of real-time systems using utilization bounds

We have developed a new technique for determining at design time, the utilization bound for a specific set of hard real-time periodic tasks with known periods, when scheduled by any arbitrary fixed priority algorithm. The technique does not depend on precise knowledge of task computation times and does not require that task deadlines coincide with the arrival of the next request of the task. In this paper, we describe the technique and discuss how it can be applied to facilitate an engineering approach to the problem of task scheduling in hard real-time systems.

Finding all minimum-size separating vertex sets in a graph

We present a new algorithm based upon network flows for finding all minimum-size separating vertex sets in an undirected and unweighted graph. The sequential implementation of our algorithm runs in Θ(Mn + C) = O(2kn3) time, where M is the number of minimum-size separating vertex sets of the graph; n, the number of the vertices in the graph; m, the number of the edges in the graph; k, the connectivity of the graph, and C = kn min(k(m + n), A), where A is the complexity of the best maximum flow algorithm for unit networks. The parallel implementation runs either in O(k log n) deterministic time or in O(log2n) randomized time using Θ(;M2n2 + knNα) = O(4k(n6/k2)) processors on a PRAM, where Nα is the number of processors needed for parallel matrix multiplication in O(log n) time on PRAM.

Substructure Allocation in Recursive Interconnection Networks

In a multiuser message passing MIMD sys tem, substructure allocation is an important aspect of sys tem design. In this paper, we present four substructure al location algorithms for a multiuser WK-Recursive network. Two algorithms are bit-map based and two are tree based. The algorithms are compared using simulation.

A note on approximating graph genus

Abstract This note, using simple combinatorial analysis, shows two interesting facts on the approximability of graph minimum genus embeddings: 1. (1) for any function ƒ(n) = O(n e ) , 0 ⩽ e G of n vertices into a surface of genus bounded by γ min (G) + ƒ(n) , unless P = NP ; and 2. (2) there is a linear time algorithm that embeds a graph G of n vertices into a surface of genus bounded by max {4 γ min ( G ), γ min ( G ) + 4 n }, where γ min ( G ) denotes the minimum genus of the graph G . An approximation algorithm with approximation ratio O(√ n ) for bounded degree graph embeddings is also presented.

Reinventing the wheel: an optimal data structure for connectivity queries

We show that, for any fixed k, there exists an optimal O(n)-space compact representation of a k-connected graph G with n vertices, such that one can determine in O(1) time whether two vertices areconnectedbyk+l vertex-dkjoint paths, or are separated by k vertices/edges. Previously, the existence of such compact representations was known only for k <3. 1 Summary of Results A fundamental issue for the fault-tolerance and reliabilityofnetworksis determining the existence of multiple disjoint paths connecting two nodes. In this paper we investigate the problem of constructing a compact representation of a graph so that one can test quickly for the existence of such paths. *Research supported in part by the National ScienceFoundation under grant CCR9007851, by the U.S. Army Research Office under grant DAAL03-91-G-o035, by the Office of Naval Research and the Defense Advanced Research Projects Agency under contract NOO014-91-J-4052, ARPA order 8225, by the NATO Scientific Affairs Division under collaborative research grant 911016, by the Pro get t o Finalimato Sisterni Informatici e Calcolo ParaUelo of the Italian National Research Council, and by the Esprit II BRA of the European Community (project ALCOM). t Department of Computer Science, University of Newcastle, Callaghan, New South Wales 2308, Australia. The work of this author was carried out in part while visiting the United States Coast Guard Academy and Brown University. t Dip=tirnento & Inforrnatica e %ternist.ica, Universit& & Roma ‘[La Sapienza”, Via Salaria 113, 00198 Rome, Italy, 5Department of Computer Science, Texas A&M University, College Station, TX 77843-3112. f Depmtment of Computer Science, Brown University, providence, RI 02912–1910. Permission to copy without fee all or part of this material is granted provided that the copies are not made or distributed for direct commercial advantage, the ACM copyright notice and the title of the publication and its date appear, and notice is given that copying is by permission of the Association for Computing Machinery. To copy otherwise, or to republish, requires a fee and/or specific permission. 25th ACM STOC ‘93-5193 /CA, USA 9 1993 ACM 0-89791 -591 -7/93 /0005 /0794 . ..