In this paper, we design and analyze a soft handoff algorithm for code-division multiple access (CDMA) systems in a Rayleigh fading environment. For each mobile, this algorithm selects a set of base stations to be in handoff with the mobile based on the average channel conditions between the mobile and the base stations. For a given target value for the average number of base stations in handoff, our handoff algorithm is optimal in that it maximizes the effective channel gain seen by mobiles. We consider both power control and rate control in the forward and reverse links of CDMA systems, where maximizing the effective channel gain minimizes the average transmit power for the power controlled case and maximizes the average received power for the rate controlled case, respectively.

Power control is a key air-interface management mechanism for any multiple-access CDMA-based wireless air-interface. Typically it has been used to ensure a target packet-error-rate (PER) for any user connection. In the presence of hybrid-ARQ (HARQ) on the reverse link (as in 3GPP2's EVDO-Rev A), there is additional flexibility that a power control implementation can utilize to ensure quality-of-service (QoS) for any connection. This paper discusses a general framework in which this can be done. It further provides an easily implement able algorithm within this framework, and substantiates its efficacy with simulations.

In this paper, we study the robustness of scheduling against channel variations. Specifically, we consider a class of generalized proportional fairness (PF) schedulers that achieve optimal throughput-fairness trade-off and study their robustness in terms of maintaining a certain fairness criterion when the channel statistics change from the additive white Gaussian noise (AWGN) to Rayleigh fading for all users. We conclude that the scheduler is robust only when it is configured to achieve the proportional fairness. For the other scheduler configurations, which are not robust, we show how fairness deviates from the ideal one when the channel statistics change and show how scheduler parameters need to be adjusted to restore the desired fairness. We also show simulation results.

Ever since the first dial-up modems appeared in the 1960s, their obsolescence has been repeatedly predicted. However, contrary to such predictions, dial-up modems thrived in the 1980s and 1990s as a result of the slow rollout of residential digital services and the unprecedented growth of internet and remote access. Since the first 300 b/s dial-up modem standard (V.21), modem speeds have increased steadily. Most recently, International Telecommunications Union (ITU) Recommendation V.90 (1998) takes advantage of the direct digital-network connection of an internet service provider (ISP) remote-access server to achieve speeds of more than 50 kb/s downstream (from ISP to a user). However, for upstream transmission (from a user to ISP), V.90 employs the older V.34 modulation (1994), which typically delivers on the order of 30 kb/s. A new ITU modem standard called V.92 increases upstream rates to above 40 kb/s, again by taking advantage of pulse code modulation connections. In this paper, we present the transmission scheme that has been adopted for V.92. It involves a generalization of Tomlinson-Harashima precoding. We predict that V.92 will be the last dial-up modem standard. However, we have to wonder whether we might be falling into the same trap into which many others have fallen in the past. The future will be the judge!.

We consider initial timing acquisition in discrete-sequence code-division multiple-access (DS-CDMA) when propagation is affected by multipath and fading, and where the base-station broadcasts a synchronization pilot signal in the form of bursts of modulated chips transmitted periodically and separated by long silent intervals. Subject to certain simplifying assumptions, we derive the maximum-likelihood (ML) estimator by solving a constrained maximization problem. Our ML timing estimator has constant complexity per observation sample. The relation to other estimation methods is addressed, and performance comparisons are provided by simulation. The proposed estimator yields good performance independently of the multipath-intensity profile of the channel, provided that the delay spread is not larger than a given maximum spread. Moreover, our estimator is fairly robust to the mismatch in the fading Doppler spectrum and provides good performance for both fast and slow fading.

We consider timing initial acquisition in DS-CDMA when propagation is affected by multipath and fading and where the base-station broadcasts a synchronization pilot signal in the form of bursts of modulated chips transmitted periodically and separated by long silent intervals. Subject to certain simplifying assumptions we derive the maximum likelihood (ML) estimator by solving a constrained maximization problem by applying the Karush-Kuhn-Tucker conditions. Our ML timing estimator has constant complexity per observation sample. The relation to other estimation methods is addressed, and performance comparisons both in terms of estimation mean-square error and cumulative density function of the channel energy falling in a certain window are provided by simulation. The resulting estimator offers good results over all fading channels with delay spread not larger than a given maximum spread.

This work considers the achievable performance for coded systems adapted to a multipath block-fading channel model. This is a particularly useful model for analyzing mobile-radio systems which employ techniques such as slow frequency-hopping under stringent time-delay or bandwidth constraints for slowly time-varying channels. In such systems, coded information is transmitted over a small number of fading channels in order to achieve diversity. Bounds on the achievable performance due to coding are derived using information-theoretic techniques. It is shown that high diversity can be achieved using relatively simple codes as long as very high spectral efficiency is not required. Examples of simple block codes and carefully chosen trellis codes are given which yield, in some cases, performances approaching the information-theoretic bounds.

We study the system capacity of cellular systems with time-division multiple access, slow time-frequency hopping (F-TDMA), and conventional single-user processing at the receivers. System capacity is formally defined as the maximum of the product of the number of users per cell times the user spectral efficiency for a given maximum outage probability. We adopt an information-theoretic definition of outage as the event that the mutual information of the block-interference channel resulting from a finite number of signal bursts spanned by the transmission of a user code word falls below the actual code rate, because of fading, shadowing, and interference. Starting from this definition, we develop a general framework which naturally takes into account many different aspects of F-TDMA cellular systems like channel reuse, channel utilization, waveform design, time-frequency hopping, voice activity exploitation, handoff, and power control strategies. Most importantly, our analysis does not rely on the choice of a particular coding scheme and can be applied to a very large class of systems in order to find guidelines for capacity-maximizing system design. A numerical example based on a typical urban mobile environment shows that there is a considerable capacity gap between actual F-TDMA systems and the limits predicted by our analysis. However, this gap can be filled by carefully designed (practical) systems, which make use of conventional single-user processing and simple coded modulation schemes.

This work considers the achievable diversity for coded systems appropriately characterized by a block-fading channel model. We are primarily interested in cases where the number of uncorrelated fading channel realizations (blocks) F, is small, so that ideal interleaving assumptions do not hold. This is usually the case in mobile radio systems which employ coded slow frequency-hopping such as the GSM system and its derivatives, We show that the diversity order is limited to a value less than or equal to F which depends on the code rate and the size of the signaling constellation. We report on the results of code searches for rate 1/n convolutional codes for simple AM constellations, which show the minimum complexity needed to achieve maximum diversity. We also present computer simulations of some codes in order to determine the effect of code complexity on the frame and bit error-rate performance.

We introduce a traffic model for circuit-switched all optical networks which we then use to calculate the blocking probability along a path for networks with and without wavelength changers. We investigate the effects of path length, switch size, and interference length (the expected number of hops shared by two sessions which share at least one hop) on blocking probability and the ability of wavelength changers to improve performance. Our model correctly predicts unobvious qualitative behaviour demonstrated in simulations by other authors.

Intersymbol interference (ISI) in direct detection optical systems can limit channel spacing in frequency division multiplexing systems, and data rates, and transmission distances in long-haul transmission. It is desirable to reduce or cancel ISI in these situations, We investigate zero-forcing electrical filters to cancel ISI, and obtain tight performance bounds for minimum noise variance zero-forcing filters. We apply the results to Mach-Zehnder and Fabry-Perot filters, and to fibers with second order dispersion. We compare the performance of zero-forcing filters to that of rectangular impulse response filters and find that zero-forcing filters are advantageous in severe ISI situations with multilevel signalling.

We consider an optical communication system where a reference signal is transmitted along with a phase modulated information carrying signal. This system has the potential of alleviating the effect of phase noise considerably. We perform a joint optimization of the receiver bandwidth and the powers of the two signals. The scheme achieves a performance that is close to that of ideal phase modulation systems when the ratio of bit rate to laser linewidth is large, and a performance that is identical to that of frequency modulation when the ratio is small. We also introduce a centered filter structure which achieves a far better performance than traditional filters due to increased the phase coherence.>

Studies the space-time characteristics of ALOHA multiple-access protocols in bidirectional bus networks where transmissions are in the form of packets of constant length. For point-to-point communications, the maximum throughput of unslotted ALOHA is known to be 1/(2e), independent of station configuration. The authors show that, with a uniform probabilistic station configuration, the maximum throughput of slotted ALOHA tends to a nonzero constant that is less than 1/(2e), when a, the end-to-end propagation delay normalized with respect to the packet transmission time, tends to infinity. However, when N stations are evenly spaced on the bus, the maximum throughput of slotted ALOHA vanishes as a tends to infinity. For broadcast communications, the maximum throughput of slotted ALOHA is well known to be 1.>

Buffer allocation for frame reassembly and queueing in the destination node of an ATM subnetwork supporting multiple frame relay connections is examined. The finite peak rate of a frame relay connection is explicitly modeled and shown to be an important factor in buffer sizing. Specifically, the average number of buffers required is shown to be large when many low-peak-rate frame relay connections are multiplexed for delivery to a single high-speed frame relay interface port, but the tail of the distribution is determined only by aggregate traffic statistics.>

The authors consider optical networks using wavelength division multiplexing, where the path a signal takes is determined by the network switches, the wavelength of the signal, and the location the signal originated. Therefore, a signal is routed through a combination of circuit switching and wavelength routing (assigning it a wavelength). They present a bound on the minimum number of wavelengths needed based on the connectivity requirements of the users and the number of switching states. In addition, they present a lower bound on the number of switching states in a network using a combination of circuit switching, wavelength routing, and frequency changing. The bounds hold for all networks with switches, wavelength routing, and wavelength changing devices. Several examples are presented including a network with near optimal wavelength re-use.>

Routing subject to multiple quality-of-service (QOS) constraints is considered. Such a problem arises in both private corporate backbone networks and public switched networks, and will become even more prevalent in the future with the emergence of asynchronous transfer mode (ATM) networks. A rule-based call-by-call source routing strategy that makes use of routing fallbacks to accommodate users with diverse QOS requirements is proposed.>

Different physical aspects that should be taken into consideration when transmitting packets in a slotted star network using wavelength-division multiplexing (WDM) are discussed. It is shown that, due to the effect of the dispersion, some padding is necessary as well as a limit on the size of the network. A method for synchronizing the transmission of slots on the different wavelengths is presented. This method is almost insensitive to the relative lengths of the transmit and receive fibers.>

A media-access protocol for high-speed packet-switched multichannel networks that are based on a broadcast topology, such as optical passive star networks using wavelength-division multiple access, is described. The protocol supports connection-oriented traffic with or without bandwidth reservation, as well as datagram traffic to integrate transport-layer functions with the media-access layer. It uses the bandwidth efficiently while keeping the processing requirements low by requiring stations to compute their transmission and reception schedules only at the start and end of each connection. Analysis results show that low blocking probabilities for connections and high network throughput can be achieved.>

The problem of transporting constant-bit-rate (CBR) traffic through a packet network is analyzed. In the system considered, CBR traffic is packetized and packets from several similar sources are multiplexed on a transmission link. The bit streams are recreated at the receiving end by demultiplexing the packets and then playing out the packets of each CBR stream. Traffic fluctuations may cause gaps to appear in the playout process. Their frequency can be reduced by adding a smoothing delay to each stream. The queueing system analyzed has periodic arrivals and deterministic service times. A method of analysis, based on the ballot theorems of Takacs (1967), is presented to provide steady-state delay distributions as well as a transient analysis of the system to predict the statistics of the time for a gap to develop in the CBR stream as a function of the smoothing delay.>

A new access protocol that handles isochronous traffic in a much better way than previous protocols for wavelength division multiplexing (WDM) star networks is presented. This protocol allows the stations to reserve part of the bandwidth for their isochronous traffic while the rest is used for asynchronous traffic. In addition, this technique requires only one transmitter and one receiver per station and is free from the constraint of a control channel.>

This paper proposes a new media-access protocol for high-speed packet-switched multichannel networks based on a broadcast topology, for example, optical passive star networks using wavelength-division multiple access. The protocol supports connection-oriented traffic with or without bandwidth reservation as well as datagram traffic, in an attempt to integrate transport-layer functions with the media-access layer. It utilizes the bandwidth efficiently while keeping the processing requirements low by requiring stations to compute their transmission and reception schedules only at the start and end of each connection. A simple analysis shows that we can achieve low blocking probabilities for connections, as well as high network throughput.

The performance of the slotted ALOHA multiple access protocol in a high-speed bidirectional passive bus network, where transmissions are in the form of packets of constant length, is studied. Slotted ALOHA is generally considered to have better throughput performance than unslotted ALOHA, whose maximum throughput is known to be 1/(2e), independent of station configuration. It is shown that, with a probabilistic station configuration, the throughput of slotted ALOHA can degrade to 1/(3e) when the end-to-end propagation delay is significantly larger than the packet transmission time. Nevertheless, in some very high speed bidirectional bus networks with a deterministic station configuration, the throughput of slotted ALOHA can be as high as 1/2.>

The authors give a distributed algorithm to compute shortest paths in a network with changing topology. The authors analyze its behavior. The proof of correctness is discussed. It does not suffer from the routing table looping behavior associated with the Ford-Bellman distributed shortest path algorithm although it uses truly distributed processing. Its time and message complexities are evaluated. Comparisons with other methods are given.>

A number of distributed network protocols for reliable data transmission. connectivity test, shortest path, and topology broadcast have been proposed with claims that they operate correctly in the face of changing topology, without the need for unbounded numbers to identify different runs of the algorithms. It is shown that the protocols do not possess all the claimed properties and that they may not terminate, because they all share a basic flaw.>

Since Fabry-Perot (FP) filters are major candidates for use as demultiplexers in wavelength division multiple access (WDMA) networks, the authors have analyzed, in an exact and unified manner, the crosstalk degradation for several different variations of the basic FP filter, compared their performances, and optimized their design parameters. A description is given of the system model (a wavelength division multiple access system consisting of M connected transmitters and receivers with each transmitter consisting of an on-off modulated fixed-frequency laser, with the 'off' power level equal to zero) and the parameters used throughout, along with a brief discussion of relevant Fabry-Perot equations and terms and the four different filter structures analyzed. The crosstalk is examined for the cases of a single-cavity FP filter, a double-pass FP filter, a two-stage double-cavity FP filter, and a vernier double-cavity FP filter. The criteria for the performance comparison were the worst-case crosstalk, BER, and crosstalk power penalty.>

A system is considered in which V users are competing for the transmission capacity of a link. The users generate messages in a Poisson manner. The message length distribution of each user is arbitrary and may differ for different users. The objective is to investigate nonpreemptive service-time independent scheduling as a means of selectively controlling the average waiting time of the users. The average waiting time assignments that can be realized are characterized. They can be used to establish, in O(V log V) computations, whether a given average waiting time assignment is feasible. The proof of the result relies on a universal scheduling strategy which is simple, is time-invariant, and can be used to realize any feasible average waiting time assignment. A waiting time cost function is associated with each user in order to investigate the problem of finding a nonpreemptive scheduling strategy that minimizes the overall waiting time cost. A set of optimality conditions is given for this problem, and an algorithm is constructed solving it in O(V) log V steps. With a simple modification, the algorithm also solves the problem of finding a nonpreemptive scheduling strategy that minimizes the lexicographic ordering of the waiting time costs. Results are extended to the preemptive case.>

For pt.I see ibid., vol.38, p.2049-51 (Nov. 1990). Priority scheduling as a means of selectively controlling the average delay of the users in a virtual circuit integrated services network is investigated. Specifically, quantifying the end-to-end delay preferences of the users on an individual basis through associated delay cost functions, the selection of the scheduling strategies on the links of the network is formulated as a convex optimization problem whose objective consists of minimizing the overall delay cost. A set of optimality conditions is given for this problem, and two distributed algorithms for solving it are constructed. The first algorithm always converges to an optimal solution, while the second algorithm is approximate. However, the second algorithm requires less coordination than the first, and by appropriately selecting the parameters the solution can be brought as close to optimality as desired. With simple modifications, the algorithms also solve the problem of selecting the scheduling strategies on the links so as to minimize the lexicographic ordering of the cost vector.>

A queuing system with infinitely many servers, and with the following queuing discipline is considered: For any two jobs i and j in the system, such that i arrived later than j , there is a fixed probability p that i will have to wait for j 's execution to terminate before i starts executing. This queuing system is a very simple model for database concurrency control via “static” locking, as well as of parallel execution of programs consisting of several interdependent processes. The problem of determining the maximum arrival rate (as a function of p ) that can be sustained before this system becomes unstable is studied. It is shown that this rate is inversely proportional to p , and close upper and lower bounds on the constant for the case of deterministic departures are found. The result suggests that the degree of multiprogramming of multiuser databases, or the level of parallelism of concurrent programs, is inversely proportional to the probability of conflict, and that the constant is small and known within a factor of 2. The technique used involves the computation of certain asymptotic parameters of a random infinite directed acyclic graph (dag) that seem of interest by themselves.

We consider some access protocols for Aloha type multiaccess channels. We argue, and show in an important case, that they can be modified to allow new transmitters to join the system at arbitrary times. This feature, known as "limited sensing" or "continuous entry," need not reduce throughput performances. In the case presented, the modified algorithm is also robust with respect to feedback errors.

A discrete time multiaccess channel is considered where the outcome of a transmission is either "idle," "success," or "collision," depending on the number of users transmitting simultaneously. Messages involved in a "collision" must be retransmitted. An efficient access allocation policy is developed for the case where infinitely many sources generate traffic in a Poisson manner and can all observe the outcomes of the previous transmissions. Its rate of success is 0.48776. Modifications are presented for the cases where the transmission times depend on the transmission outcomes and where observations are noisy.

It is shown that the capacity region of the asynchronous multiple-access channel differs from that of the synchronous channel only by the lack of a convex hull operation.

A distributed algorithm is presented for constructing minimum weight directed spanning trees (arborescences), each with a distinct root node, in a strongly connected directed graph. A processor exists at each node. Given the weights and origins of the edges incoming to their nodes, the processors follow the algorithm and exchange messages with their neighbors until all arborescences are constructed. The amount of information exchanged and the time to completion areO(|N|^{2}).

A distributed algorithm is presented that constructs the minimum-weight spanning tree in a connected undirected graph with distinct edge weights.A processor exists at each node of the graph, knowing initially only the weights of the adjacent edges.The processors obey the same algorithm and exchange messages with neighbors until the tree is constructed.The total number of messages required for a graph of N nodes and E edges is at most 5N log2N + 2E, and a message contains at most one edge weight plus log28N bits.The algorithm can be initiated spontaneously at any node or at any subset of nodes.

An algorithm is given to find a prefix condition code that minimizes the value of the moment generating function of the codeword length distribution for a given positive argument. This algorithm is used in an iterative way to yield a code that maximizes the rate of decay of the probability of buffer overflow as the buffer length increases.

Letp(i)be a probability measure on the set of nonnegative integers. The Huffman optimum encoding technique is extended to a class ofp(i)including those whose tail decreases