Joseph L. Hammond

Properties of a transmission assignment algorithm for multiple-hop packet radio networks

We study an algorithm, originally defined by Lyui (1991), for assigning collision-free broadcast transmissions for packet radio networks. The method has a number of advantages - it is distributed, fair, permits each terminal to calculate a different frame size, requires only a limited exchange of information, and does not require a contention-based channel-access protocol. These properties make the assignment algorithm particularly useful for dynamic ad hoc networks because the transmission assignments can be easily and quickly adapted to local changes in the topology. An improved description of Lyuis algorithm is given and new simulation results are presented to demonstrate the properties of the transmission assignments for a variety of large networks.

Simulation methods for studying nonstationary behavior of computer networks

Methods for the simulation of the nonstationary behavior of computer networks are discussed. Steady-state performance measures, typically computed over time histories, are modified to be applicable to the nonstationary case by redefining the measures as ensemble statistics. The application of well-known ensemble simulation techniques to determining the modified performance metrics from finite ensembles generated by independent replications is given. The simulation methods are illustrated with a comprehensive tutorial example and with a performance study which gives results for a shared buffer switch. >

Distributed and adaptive TDMA algorithms for multiple-hop mobile networks

Time-division multiple-access (TDMA) protocols can support contention-free link-level broadcast transmissions in mobile, multiple-hop wireless networks. However, it is typically difficult to adapt TDMA assignments to dynamic topologies. The paper defines and discusses two algorithms for adjusting TDMA slot assignments to the dynamic changes of a mobile network. Our new protocols offer three significant advantages: (1) the information required by the protocols is gathered exclusively from a terminals local neighborhood; (2) global coordination is not required; (3) changes to the state information at a terminal do not need to be distributed to all terminals in the network. The operation of the algorithms is evaluated with an extensive simulation study using a large number of randomly generated networks in which the diameters, degrees, and number of nodes are varied. Results show that the efficiency of the algorithms, determined as the average number of transmission slots assigned per terminal, is better under static conditions than two cases chosen for reference; the efficiency does not degrade significantly under dynamic conditions for which the algorithms continually adjust to added and deleted links. For the networks considered, the algorithms typically converge in times on the order of several hundred slots and a convergence bound approximately four times larger is established. Convergence time is shown to increase in almost a linear fashion with terminal radius and it is relatively insensitive to the number of terminals.

An Efficient Simulation Technique for Performance Studies of CSMA/CD Local Network

A new simulation technique is presented for performance studies of local area networks which use CSMA/CD as the access protocol. The method requires partitioning the network stations into a few primary stations and the remainder as background stations. The background stations are modeled as a group by a background algorithm based on a dynamic CSMA/CD model which uses a set of probabilities and sampling to produce, in a recurrent manner, a stochastic sequence of busy/idle periods. The new method is more efficient than conventional discrete-event simulations when the number of network stations is reasonably large. The algorithm and the overall simulation technique are comprehensively validated. The technique is illustrated with a performance study of interactive transactions on a metropolitan CATV network.

An adaptive transmission-scheduling protocol for mobile ad hoc networks

Transmission-scheduling protocols can support contention-free link-level broadcast transmissions and delay sensitive traffic in mobile, multiple-hop packet radio networks. Use of transmission-scheduling protocols, however, can be very inefficient in mobile environments due to the difficulty in adapting transmission schedules. The paper defines a new adaptive and distributed protocol that permits a terminal to adapt transmission assignments to changes in topology using information it collects from its local neighborhood only. Because global coordination among all the terminals is not required and changes to transmission assignments are distributed to nearby terminals only, the protocol can adapt quickly to changes in the network connectivity. The two key parameters that affect the ability of the protocol to adapt to changes in connectivity are the rate of connectivity changes and the number of terminals near the connectivity changes. Using simulation, we determine the ranges for these parameters for which our adaptive protocol can maintain collision-free schedules with an acceptable level of overhead. The stability of the protocol is also characterized by showing that the protocol can quickly return to a collision-free transmission schedule after a period of very rapid changes in connectivity. Our channel-access protocol does not require a contention-based random-access phase to adapt the transmission schedules, and thus its ability to adapt quickly does not deteriorate with an increase in the traffic load.

A distributed load-based transmission scheduling protocol for wireless ad hoc networks

This paper presents a load-based transmission scheduling (LoBaTS) protocol for wireless ad hoc networks. Since terminals in these networks may be required to forward unequal amounts of traffic, a transmission scheduling protocol which assigns equal transmission capacity to all terminals may result in traffic bottlenecks as traffic load increases. The LoBaTS protocol allows terminals to alter their existing transmission schedules so that those terminals which are required to forward more traffic can reserve additional transmission slots, thereby alleviating traffic bottlenecks. The LoBaTS protocol is designed to avoid creation of new traffic bottlenecks, maintain collision-free broadcast transmissions, and execute with minimal disruption to the network. In addition, it requires only local information to operate, making it suitable for implementation in a distributed-control environment. Simulations utilizing a variety of test networks and traffic conditions show that the LoBaTS protocol greatly improves end-to-end packet delay, throughput, and completion rate when compared to a transmission schedule which does not account for traffic load.

The performance of adaptive window flow controls in a dynamic load environment

The behavior of window flow control schemes that adapt to changing network conditions is studied. A dynamic window scheme, which adjusts the window size based on explicit network congestion indicators, is proposed. The throughput performance of the adaptive policy under both steady-state and nonstationary conditions is studied. The superiority of this scheme over some previously proposed adaptive window schemes based on implicit congestion detection is demonstrated.<<ETX>>

An analysis of the congestion effects of link failures in wide area networks

The results of a study to determine the effects of link failures on network performance are presented. The network studied is a virtual-circuit-based packet-switched wide area network. A generic queuing framework is developed to study the effect of failures, and the subsequent traffic restoration, on network performance. In general, the congestion resulting after a failure is a transient phenomenon. Hence, a numerical-method-based nonstationary queuing analysis is conducted in order to quantify the effects of failures in terms of the transient behavior of queue lengths and packet loss probabilities. A bounding relationship is developed whereby a network node can determine whether or not congestion will occur as the result of traffic restoration after a failure.<<ETX>>

Distributed formation of broadcast transmission schedules for mobile ad hoc networks

In a mobile ad hoc packet radio network using transmission scheduling, when a terminal boots and is ready to join the network, it has no knowledge of the networks current state. In order to send and receive packets without interference, the new terminal must have some means of notifying other terminals of its presence and exchanging information with them so that it can form a collision free broadcast transmission schedule. We demonstrate the performance of a distributed protocol designed to allow a new terminal to acquire sufficient information about its local environment to achieve collision free transmissions. Furthermore, we show that changes to the network are limited to the new terminals local neighborhood. As a result, the size of the overall network has very little effect on the performance of the protocol. The protocol is not only effective in adding a new terminal; it can also organize a group of terminals into a new network. This is demonstrated for the extreme case of all terminals powering up simultaneously. The overall performance of the algorithm depends primarily on the density of the network and rate at which changes occur in the network.

Evaluation of a split-connection mobile transport protocol

A performance evaluation is presented for a split-connection protocol for wireless Internet access which is denoted Split-Connection Mobile Tranport Protocol (SCMTP). It uses the general approach of the previously introduced Mobile End Transport Protocol (METP) but with differences that include a wireless-link channel-access protocol better matched with current cellular networks and more general ARQ methods for error control in the wireless link. In common with METP, SCMTP uses a standard TCP protocol on the wire-line connection and isolates the data flow in the wire-line network from the effect of wireless packet errors.Performance is considered for the important case of a single SCMTP split connection between a fixed host and a mobile host with heavy downlink traffic from the fixed host to the mobile host. It is shown for these conditions that if the wire-line packet-error rate is small, a steady state is reached in which the connections data flow in the wire-line network remains under the control of the receiver-advertised window of the TCP entity at the base station. Performance is evaluated for the steady-state operation of the SCMTP protocol, and relationships are established between the key properties of the split connection and the end-to-end performance of the connection.It is shown that for heavy downlink traffic, the delay in the wire-line part of the connection does not affect steady-state throughput if the receiver buffer is sized appropriately. It is also shown that use of the go-back-N ARQ protocol on the wireless link yields better performance than the stop-and-wait ARQ protocol, although the performance with go-back-N ARQ is more sensitive to the characteristics of the wireless channel. It is shown that under a broad range of conditions, SCMTP with go-back-N ARQ provides nearly optimal utilization of the capacity of the wireless link.