Abstract: A priority queuing system m which a customer can feed back and change priority after being served is analyzed. More exactly, there are c customer types. Customer type i feeds back after service N(I) 1 times. The kth time a type / customer is m the queue it has priority level f(t, k) and requires a service time with arbitrary distribution G,k Type ~ customers enter from outside as a Poisson process w~th rate X, and can be e~ther preemptive or nonpreemptlve. A customer who has been preempted must start serwce over from the beginning w~th a new (independently chosen) service Ume. The analysis gwes mean waiting times for each customer type at each stage m its itinerary, the mean number of each customer type m the system, and other quantities of interest. It ~s also shown that bulk arrivals and branching can be included m the model without substantially changing the form of the solution.

Abstract: Consider a closed cyclic queuing system consisting of M exponential queues. The Laplace-Stieltjes transform of the joint distribution of the consecutive sojourn times of a customer at the M queues is determined and shown to have a product form. The proof is based on a reversibility argument.

Abstract: We study the problem of optimal customer admission to multiserver queues. These queues are assumed to live in an extraneous environment which changes in a semi-Markovian way. Arrivals, service mechanism and random reward/cost structure may all depend on these surroundings. Included as special cases are SM/M/c queues, in particular G/M/c queues, in a random environment. By a direct inductive approach we establish optimality of a generalized control-limit rule depending on the actual environment. Particular emphasis is laid on different applications that show the versatility of the proposed setup.

Abstract: The FETCH-AND-ADD instruction provides for synchronization of multiple processes in a parallel manner. This paper explores the use of FETCH-AND-ADD in the context of database systems. We show how to enqueue locks, detect lock conflicts, and release locks without resorting to critical program sections that require mutual exclusion during execution. The scheme is compatible with a variant of lock management proposed by Rosenkrantz and Stearns. A second approach to parallel lock management is based on a reservation scheme by Milenkovic. This methodology uses FETCH-AND-ADD implementation of a priority queue. An implementation of such a queue originally reported by Gottlieb and Kruskal is used for this purpose, although the storage requirements for queue management may be unacceptably large in specific cases. Both approaches described in the paper suggest that FETCH-AND-ADD is potentially effective for eliminating serial bottlenecks caused by lock conflicts in multiprocessor systems.

Abstract: A telephone switching office having essential service protection is disclosed. As each service request is received it is put in a common queue. Periodically, and regardless of the load, the common queue is examined for requests that have not been served with a given time interval. If these requests are from essential lines, they are put in a special queue. When a server, such as a digit receiver or processor is available the special queue is served prior to the common queue.

Abstract: Fair and fast service to a plurality of channels served by packet switches at nodes interconnected through a common carrier facility is achieved by controlling the quantity of data sent from any one channel to another. When data arrives at a node from a channel, the channel address is entered in line in the first of two channel address queues and the data is entered in a separate data queue. Each channel address is read from the first queue, a predetermined number of data bytes is read from the corresponding data queue and sent in a frame to a distant node. Thereafter, the channel address is entered at the end of the second, lower priority queue. The frame may have data bytes from other first queue channels but the frame length is limited either by a predetermined number of bytes or when no more channel addresses remain in the first queue. When no channel addresses remain in the first queue, a new frame is started by reading a channel address from the lower priority second queue. A frame may have data from a data queue corresponding to a channel address in the second queue followed by data from a data queue corresponding to a channel in the higher priority first queue.

Abstract: We consider n independant identical deterministic packet streams. Packets must share the transmission capacity of a single outgoing trunk.

Abstract: A microprocessor with a macro-rom exhibits reduced latency time and greater flexibility by including both a macro-rom queue and a main program queue. The arrangement eliminates the undesirable latency associated with fetching program as part of a return sequence from a macro-rom instruction. Also, the arrangement allows parameters to be extracted from the main program queue as the macrosequence is executing from the macro-roms program queue.

Abstract: We study a scheduling problem on a finite-source queueing model which is identical to a two-stage closed Markovian cyclic queue consisting of a single-server "processor" and an infinite-server "terminals"; each job has a distinct mean service time at the processor. We obtain a necessary and sufficient condition for a performance vector consisting of utilization factors or average response times of the processor for the jobs to be achievable by some scheduling strategy. The result is analogous to that obtained by Coffman and Mitrani on infinite-source single-server queues.

Abstract: : This note describes the Dyna-Sim model, which provides a capability for exploring the implications of maintenance repair queing and requirements for spare parts Several useful conclusions emerge from the study of the Automated Test Equipment (ATE) queing problem using Dyna-Sim, which are summarized as follows: The ample server assumption (more than enough ATE available to serve any repair demands) is a very poor approximation when queues become saturated in high aircraft sortie rate scenarios; and if repair times follow the exponential distribution, certain approximation techniques become available (because the queuing system satisfies requirements of a Markov process) But real-world repair times are rarely exponential The authors used Dyna-Sim to show that in a constrained server problem, the choice of the repair time distribution is not important Thus, analytical approximations for queuing in capability assessment models are traceable

Abstract: In a system of single-server queues in series, a customer enters the first queue ; he waits till he is served, enters the second queue, and so on. A heuristic method of approximating the distribution function of the inter-departure time through erlangian distribution is presented in this paper. It is proved that if all the queues are in equilibrium, the expected value of the inter-departure time at each of the queues in tandem equals the expected value of the inter-arrival time in the first queue. For a queue not in equilibrium, the inter-departure time follows asymptotically the distribution of its service time. The output from an unstable queue equals the input into the next queue in tandem if the latter is in equilibrium. The approach given in this paper is supported though simulation experiments.

Abstract: It has been our experience that, when writing software for operating systems, the synchronization techniques available in Communicating Sequential Processes are inconveniently primitive. In this paper we investigate how different types of queue can be represented in C.S.P., and we show how this approach can be developed into a general queuing mechanism.

Abstract: A popular means of increasing the effective rate of main storage accesses in a large computer is a multiplicity of memory modules accessible in parallel. Although such an organization usually achieves a net gain in access rate, it also creates new modes of congestion at the storage controller. This paper analyzes the variables that describe such a congestion: queue lengths and delays. A controller that maintains separate register sets to accommodate the request queue of each module is considered. The various processors attached to the storage are assumed to generate, in each memory cycle, a number of access requests with the same given distribution. The addresses specified by these requests (reduced to the module index) are further assumed to follow the states of a first-order Markov chain. The analysis then becomes one of a single-server queuing system with constant service time and indexed batch arrival process. Results are derived for several descriptors of the congestion and thus of the quality of service offered by such an organization. The aim throughout is to embody the results in a form readily suitable for numerical evaluation.

Abstract: Consider a message transmission protocol where the sender waits for acknowledgement of each message before sending the next. Obviously this scheme can be modeled by a single server FIFO queue where the clients are the messages, the queue line is the sender's buffer queue and the service time is the delay in getting the message to receiver. Now consider another protocol where the sender does not wait for ack before sending the next message. Messages could be acked selectively and will lead to selective retransmission. The receiver gets the messages in any order and buffers them up for sequential delivery to the user. Messages have to be buffered both at the sender and receiver. The buffer is not a FIFO queue as sequencing and selective retransmission spoil the order at both ends. Can we still model the protocol for total delay using a FIFO queue?

Abstract: This paper studies an M/G/1 queueing system with a finite waiting room and with server vacation times consisting of periods of time that the server is away from the queue doing additional work. This model has been used in conjunction with a related model to analyze the performance of a processor with a cyclic scheduling algorithm and where, due to finite queueing capacities, losses are a primary concern. Service at the queue is exhaustive, in that a busy period at the queue ends only when the queue is empty. At each termination of a busy period, the server takes an independent vacation. The queue length process is studied using the embedded Markov chain. Using a combination of the supplementary variable and sample biasing techniques, we derive the general queue length distribution of the time continuous process, as well as the blocking probability of the system, due to the finite waiting room in the queue. We also obtain the busy period and waiting time distributions.