Jin-Fu Chang

Analysis of ARQ protocols via signal flow graphs

A unified methodology using signal flow graphs is proposed and used to analyze several existing ARQ (automatic repeat request) protocols. The methodology offers a systematic and efficient way of obtaining throughput and delay. It also offers additional advantages in obtaining results such as the mean response time, and in analyzing the ARQ protocols in which the feedback channel is not error-free. >

End-to-end delay of an adaptive selective repeat ARQ protocol

This paper analyzes the end-to-end delay of a selective-repeat ARQ protocol which tries to adaptively use the channel idle time. The end-to-end delay consists of three parts: queueing delay at the transmitter, transmission delay over the channel, and resequencing delay at the receiver. Although this paper succeeds only in obtaining approximate end-to-end delay, the agreement with simulation is generally good. >

A Multibeam Packet Satellite Using Random Access Techniques

The performance of a multibeam packet satellite using ALOHA type random access techniques is studied. The satellite provides either no buffer or an infinite buffer for successful packets on the uplink channels. A TDMA frame is organized in accordance with the traffic demand from each area. System throughput and packet delay have been successfully obtained. Aiming at improving the performance of the system, three different protocols are introduced and studied in the zero buffer case. It is demonstrated, through examples, that satisfactory system performance can be obtained using an appropriate protocol. However, system performance can also be severely degraded if the transition time for the satellite transponder to switch from one zone to another is not negligible. In this case, the performance can be restored via the use of a buffer of sufficiently large size.

The Behavior of a Finite Queue with Batch Poisson Inputs Resulting from Message Packetization and a Synchronous Server

This paper studies the behavior of a packet switch which provides finite waiting space and receives packetized messages. The arrivals of the messages constitute a Poisson process. Each message consists of a random number of packets. The number of packets contained in a message is assumed to be an integer-valued random variable which may follow any arbitrary probability distribution. All packets residing in the buffer receive service from a single output transmitter operating synchronously at a constant rate. Each packet receives the same fixed service time from the transmitter and then leaves the system. Upon the arrival of a message, if the remaining buffer space is not enough to accommodate all packets of the message, then the entire message is completely rejected. Results such as message blocking probability, packet blocking probability, throughput, and mean delay have been obtained. Two different approaches, minislot approximation and the application of the residue theorem, are used to obtain these results. Especially, this combinatorially very complex problem is successfully solved by the residue theorem in a recursive manner. These results are useful in evaluating the performance of a packet switch. They are also useful for design purposes.

Enhancing the capacity of DS-CDMA system using hybrid spreading sequences

A new direct-sequence code-division multiple-access (DS-CDMA) system using hybrid spreading sequences in order to increase system capacity is proposed. This scheme permits accommodating more than N users, where N is the processing gain of the spreading sequences. Users are divided into two classes. The centerpiece idea is to synchronously assign N mutually orthogonal codes to class-1 users, and asynchronously assign quasi-orthogonal codes to class-2 users. The proposed architecture is realized through the DS-CDMA and multicode CDMA (MC-CDMA) techniques for class-1 and class-2 users, respectively, and the system capacity is considerably enhanced as compared with Sari et al.. The new approach still excels over Vanhaverbeke et al. when the number of class-1 users K/sub 1/ is large; but our desire is to expand system capacity. Compared with Sari et al. and Vanhaverbeke et al., the new design is able to support multirate applications, and is especially suitable for high-speed multimedia transmissions. Furthermore, the performance of the MC-CDMA system over an additive white Gaussian noise channel is evaluated for various primary codes and mutually orthogonal subcodes. It is shown that orthogonal Gold codes concatenated with Hadamard sequences yield the lowest bit-error rate when the signal-to-noise ratio lies between 0 and 30 dB.

The effect of multipath interference on the performance of packet radios

The effect of multipath interference on the performance of packet radios is investigated. The protocols considered include pure ALOHA and slotted ALOHA. Signal reception with and without capture effect is considered in the analysis. Mathematical expressions of system throughput and packet delay have been successfully obtained for each protocol. Numerical calculations are used to demonstrate when the multipath interference can be ignored and when it can be disastrous. >

The effect of prioritization on the behavior of a concentrator under an accept, otherwise reject strategy

The effect of prioritization on the behavior of a concentrator which accepts packetized messages and receives a service from a transmitter operating in synchronous mode is studied. The messages arrive according to a Poisson process, and are divided into two classes: ordinary and priority. A buffer is shared between ordinary and priority packets. Two buffer-sharing strategies are considered: sharing with maximal length, and complete sharing subject to purgation. Upon the arrival of a message, if its size exceeds the remaining buffer space, then the message will be completely rejected. For each buffer-sharing strategy, results obtained include state transition probability, state probability, blocking probability, throughput, and delay. Computer simulations are used to verify the validity of results. >

A hybrid contention-based TDMA technique for data transmissions

A hybrid contention-based time-division multiaccess (TDMA) protocol is proposed and studied. Channel time is organized into frames, each consisting of reservation, fixed, and variable subframes. Users are divided into groups. Each group is given a data slot in the fixed subframe that is open to users in that group on the basis of contention. Each user can also place his demand in the reservation subframe. The number of slots assigned to a particular user is based on the information gathered from the other two subframes. Performance analyses, both transient and steady-state, are successfully carried out. Computer simulations verify the analyses. >

Packet Satellite System with Multiuplinks and Priority Discipline

The performance of a packet satellite system with multiple uplink channels using priority service discipline for successful packets on the uplinks is analyzed in this paper. The satellite is assumed to provide either no on-board storage or an infinite buffer for successful packets on the uplinks. The delay-throughput characteristics for each priority as well as the entire system have been obtained. Expression for the z -transform of the system state at equilibrium has been derived for the case of infinite buffer. As expected, it is demonstrated through examples that packets with lower priorities have lower throughput and suffer longer delay.

Collision resolution for variable-length messages

A collision resolution algorithm for variable length messages is proposed and studied. Both throughput and delay are obtained. The delay analysis is done at both the packet and message level. Due to reduced contention, this protocol can achieve very high throughput. The effect of channel errors is also studied. >