Exercises in Computer Systems Analysis

The single server queue is one of the most basic queueing systems to model stochastic waiting dynamics. Most work involving the single server queue only analyzes the customer or agent behavior. However, in this work, we are inspired by COVID-19 applications and are interested in the interaction between customers and more specifically, the time that adjacent customers overlap in the queue. To this end, we derive a new recursion for this overlap time and study the steady state behavior of the overlap time via simulation and probabilistic analysis. We find that the overlap time between adjacent customers in the M/M/1 queue has a conditional distribution that is given by an exponential distribution, however, as the distance between customers grows, the probability of a non-negative overlap time decreases geometrically. We also find via simulation that the exponential distribution still holds when the distribution is non-exponential, hinting at a more general result.

A model for the service time structure in the single server queue is given embodying correlations between contiguous and near-contiguous service times. A number of results are derived in the case of Poisson arrivals both for equilibrium and the transient state. In particular, Kendall's (equilibrium) resultP(a departure leaves the queue empty) = 1 — (mean service time)/(mean inter-arrival time) is found still to hold good. The effect of the correlation on the mean and variance of the equilibrium queue length distribution is examined in a simple case.

It is well known that queues with exponentially distributed service times have the smallest Shannon capacity among all single-server queues with the same service rate. In this paper, we study the capacity of timing channels in which the service time distributions have bounded support, i.e., Bounded Service Timing Channels (BSTC). We derive an upper bound and two lower bounds on the capacity of such timing channels. The tightness of these bounds is investigated analytically as well as via simulations. We find that the uniform BSTC serves a role for BSTCs that is similar to what the exponential service timing channel does for the case of timing channels with unbounded service time distributions. That is, when the length of the support interval is small, the uniform BSTC has the smallest capacity among all BSTCs.

In this paper, a single-server queue with non-homogeneous Poisson input and general service time is considered. Particular attention is given to the case where the parameter of the Poisson input λ(t) is a periodic function of the time. The approach is an extension of the work of Takács and Reich . The main result of the investigation is that under certain conditions on the distribution of the service time, the form of the function λ(t) and the distribution of the waiting time at t = 0, the probability of a server being idle P 0 and the Laplace transform Ω of the waiting time are both asymptotically periodic in t. Putting where b(t) is a periodic function of time, it is shown that both P o and Ω can be expanded in a power series in z, and a method for calculating explicitly the asymptotic values of the leading terms is obtained. In many practical queueing problems, it is expected that the probability of arrivals will vary periodically. For example, in restaurants or at servicestations arrivals are more probable at rush hours than at slack periods, and rush hours are repeated day after day

This paper considers a queueing system consisting of two single-server queues in series, in which the service times of an arbitrary customer at both queues are identical. Customers arrive at the first queue according to a Poisson process. Of this model, which is of importance in modern network design, a rather complete analysis will be given. The results include necessary and sufficient conditions for stationarity of the tandem system, expressions for the joint stationary distributions of the actual waiting times at both queues and of the virtual waiting times at both queues, and explicit expressions (i.e., not in transform form) for the stationary distributions of the sojourn times and of the actual and virtual waiting times at the second queue. In Part II (pp. 644–659) these results will be used to obtain asymptotic and numerical results, which will provide more insight into the general phenomenon of tandem queueing with correlated service times at the consecutive queues.

We propose a static service differentiation policy for a single-server queueing system serving homogeneous customers. We show that by randomly assigning customers different service grades with different service rates, the average waiting time can be reduced without affecting the mean service time. Such differentiation introduces more service time variability, but it also creates information that enables the implementation of service rate-based scheduling, which mitigates the increased variance and may even reduce the total waiting time. We provide conditions under which our static service differentiation reduces waiting, and further derive closed-form expressions for the optimal differentiation policy, which shows that both optimal service rates and allocation probabilities form geometric sequences. We illustrate our policy in the context of quality-based service domains, in which customers value service time but dislike waiting. Numerically, we find that providing differentiated service can improve system performance by 5% without any additional capacity.

Single server queues with batch arrivals are useful in the modeling of computer and communication systems amongst others. In this paper we give for a wide class of single server queues with batch arrivals a stable recursive method to compute state probabilities both at arrival and at arbitrary epochs. This class includes finite capacity models with state dependent Markovian input in which batches could be totally or partially rejected. The customers of a batch are served individually and the service time has a general distribution function. In particular, we discuss as application the computation of operating characteristics, such as rejection probability and probability of server idleness, for a finite capacity M/G/1 queue for which arriving customers are rejected if their waiting plus service times exceed the capacity of the system.

A discrete-time single server queue with FIFO discipline and general independent service times is studied. The authors present a relationship between the distribution of the delay of an arbitrary customer and the joint distribution of the buffer occupancy during an arbitrary (non-empty) slot, and the remaining service time for the customer in service at the beginning of such a slot.

Consider a single server queue with renewal arrivals and i.i.d. service times in which the server operates under a processor sharing service discipline. To describe the evolution of this system, we use a measure valued process that keeps track of the residual service times of all jobs in the system at any given time. From this measure valued process, one can recover the traditional performance processes, including queue length and workload. We show that under mild assumptions, including standard heavy traffic assumptions, the (suitably rescaled) measure valued processes corresponding to a sequence of processor sharing queues converge in distribution to a measure valued diffusion process. The limiting process is characterized as the image under an appropriate lifting map, of a one-dimensional reflected Brownian motion. As an immediate consequence, one obtains a diffusion approximation for the queue length process of a processor sharing queue.

This paper considers a single server queue in which the service time is exponentially distributed and the service station may breakdown according to a Poisson process with the rates γ and γ' in busy period and idle period respectively. Repair will be performed immediately following a breakdown. The repair time is assumed to have an exponential distribution. Let g(t) and G(t) be the probability density function and the cumulative distribution function of the interarrival time respectively. When t tends to infinity, the rate of g(t)/[1 – G(t)] will tend to a constant. A set of equations will be derived for the probabilities of the queue length and the states of the arrival, repair and service processes when the queue is in a stationary state. By solving these equations, numerical results for the stationary queue length distribution can be obtained.

A significant amount of recent research has been focused on the stability of multiclass open networks of queues (MONQs). It has been shown that these networks may be unstable under various queueing disciplines even when at each one of the nodes the arrival rate is less than the service rate. Clearly, in such cases the expected delay and the expected number of customers in the system are infinite. In this paper we propose a new class of scheduling rules that can be used in multiclass queueing networks. We refer to this class as the stable shortfall-based priority (SSBP) rules. This SSBP class itself belongs to a larger class of rules, which we refer to as the shortfall-based priority (SBP) rules. SBP is a generalization of the standard non-preemptive priority rule in which customers of the same priority class are served first-come, first-served (FCFS). Rules from SBP can mimic FCFS as well as the so-called strict or head-of-the-line priority disciplines. We show that the use of any rule from the SSBP class ensures stability in a broad class of MONQs found in practice. We proceed with the construction of a sample path inequality for the work done by an SSBP server and show how this inequality can be used to derive upper bounds for the delay when service times are bounded. Bounds for the expected delay of each class of customers in an MONQ are then obtained under the assumptions that the external arrival processes have i.i.d. interarrival times, the routings are deterministic and the service times at each step of the route are bounded. In order to derive these bounds for the average delay in an MONQ we make use of some of the classical ideas of Kingman.

Conventional wisdom suggests that a fast single-server queue is preferable to a number of slower single-server queues. The rationale is that if the service rate of the superposed queue is the sum of the service rates of the individual queues, then the average waiting time for the superposed queue is smaller than the weighted average waiting time of the individual queues. This argument, however, assumes that the service time of the customers is perfectly scalable. Unfortunately, this assumption rarely holds in reality. In this paper we analyze the superposition of M/G/1 queues. We assume that there are scalable and unscalable components in the service time. On the basis of the criteria of average waiting time and average response time, we derive necessary and sufficient conditions for whether the superposition of queues is preferable. Examples of magnetic rotational disks are given to illustrate our findings.

In this paper, a single-server queue with negative customers is considered. The arrival of a negative customer will remove one positive customer that is being served, if any is present. An alternative approach will be introduced to derive a set of equations which will be solved to obtain the stationary queue length distribution. We assume that the service time distribution tends to a constant asymptotic rate when time t goes to infinity. This assumption will allow for finding the stationary queue length of queueing systems with non-exponential service time distributions. Numerical examples for gamma distributed service time with fractional value of shape parameter will be presented in which the steady-state distribution of queue length with such service time distributions may not be easily computed by most of the existing analytical methods.

A novel flexible discipline for providing priority in a single-server queue is applied to the system where the service can be unreliable what results in the loss of a customer or repetition of its service. According to this discipline, arriving customers are stored in the finite buffers dedicated to the customers of the corresponding type if the buffer is not full. After the staying in the corresponding buffer during the exponentially distributed time, the customers try to enter the main buffer of a finite capacity which is common for both types of customers. If this main buffer is full, the customer returns to the dedicated buffer and repeats the attempts to enter the main buffer later. Customers staying in the dedicated buffers are impatient and can go away from the system after a certain patience period duration of which has an exponential distribution. Customers of both types, which succeed to enter the main buffer, are picked up for the service in the order of their admission to this buffer. Providing the preference to priority customers is managed via the corresponding choice of capacities of the dedicated buffers and the rate of trials to transfer to the main buffer. Performance measures of this system are obtained under the assumption that the arrival flows of two types of customers are defined by the Markov arrival processes and the service time has the distribution of phase-type with failures type. Some aspects relating to an optimal choice of the parameters of the system are discussed via numerical experiments.

We study a discrete-time, single-server queue with partial buffer sharing. There are two priority classes of jobs. Though class 1 jobs in the queue have higher priority for the next service than any of class 2 jobs, class 2 jobs are allowed to occupy their own part of buffer when the shared part of buffer is full. We characterize a bursty arrival process using bursts which consist of the same class of jobs. Once the first job of a burst arrives at the queue, the successive jobs will arrive on every time slot until the last job of the burst arrives. The numbers of jobs of a burst and the inter-arrival times of bursts are assumed to be i.i.d., respectively, and the service time is assumed to be equal to one slot. This model targets the buffer management to meet the quality of service requirments of different traffic types as video, voice and data in ATM multiplexer. In particular, class 1 jobs may correspond to cells with the strict delay requirments. On the other hand, class 2 jobs may correspond to cells with the strict cell loss requirments. We propose an efficient numerical method to exactly obtain the job loss probability, the waiting time distribution and the mean queue length. Some numerical examples are also given.