Modeling and analysis of a communication switching processor (original) (raw)
Related papers
Dynamic scheduling for switched processing systems with substantial service-mode switching times
Queueing Systems, 2008
Switched Processing Systems (SPS) represent canonical models for many communication and computer systems. Over the years, much research has been devoted to developing the best scheduling policies to optimize the various performance metrics of interest. These policies have mostly originated from the well-known MaxWeight discipline, which at any point in time switches the system into the service mode possessing "maximal matching" with the system state (e.g., queue-length, workload, etc.). However, for simplicity it is often assumed that the switching times between service modes are "negligible"-but this proves to be impractical in some applications. In this study, we propose a new scheduling strategy (called the Dynamic Cone Policy) for SPS, which includes substantial service-mode switching times. The goal is to maximize throughput and maintain system stability under fairly mild stochastic assumptions. For practical purposes, an extended scheduling strategy (called the Practical Dynamic Cone Policy) is developed to reduce the computational complexity of the Dynamic Cone Policy and at the same time mitigate job delay. A simulation study shows that the proposed practical policy clearly outperforms another throughput-maximizing policy called BatchAdapt, both in terms of the average and the 95th percentile of job delay for various types of input traffic.
Improving Quality of Service for Switched Processing Systems
2006 11th Intenational Workshop on Computer-Aided Modeling, Analysis and Design of Communication Links and Networks, 2000
Switched Processing Systems (SPS) capture the essence of a fundamental resource allocation problem in many modern communication, computer and manufacturing systems involving heterogeneous processors and multiple classes of job traffic flows. Recently, increased attention has been paid to the issue of improving quality of service (QoS) performance in terms of delays and backlogs of the associated scheduling policies, rather than simply maximizing the system's throughput. In this study, we investigate a class of throughput maximizing scheduling policies called MaxProduct policies. The objective is that through a use of dynamically changing "optimal" queue weights, the corresponding QoS performance measure -e.g. the average system delay-can be significantly improved. The proposed approach involves a statistical smoothing technique for tracking the system's workload and utilizes the result of how the MaxProduct policies drain out an initially placed workload in the shortest possible time. It is further shown that the proposed modification of the MaxProduct policy, achieves maximum throughput without requiring knowledge of the incoming traffic's statistics. The scheduling policy is illustrated on a small SPS subject to different types of input traffic.
Message Waiting Time in a Packet Switching System
Journal of the ACM, 1980
The behavior of a buffer having an arbitrary number of common output channels in a packet switchmg system with an arbitrary number of priority classes of messages as studied Because of the random lengths of the messages and the fixed packet size, some messages must be split into several packets. Packets of the same message may not be sent immediately m sequence because of the packets of higher priority messages Therefore, the message wamng tame is defined as the wamng tame of the last packet carrying a part of the given message The hmmng probability dastnbution of thas delay is calculated for the case of mdependent packet amval processes wRh stataonary independent increments Some numerical results are also presented The results obtamed can be used m the analysis of message path delay in certain store-and-forward commumcatton systems, particularly loop communication systems. KEY WORDS AND PHRASES queuemg theory, Markov processes, packet switchmg, priority queueing dascaplines, loop communication systems CR CATEGORIES 3 81, 5 5, 8 1 1. Introduction The commumcatlon system in a computer network has an essential influence on the performance of the whole network. In order to improve the utilization of channels and decrease communication costs, the packet switching technique is frequently used. Fixed size packets are sent between the nodes of such a system; these packets are created after partitioning random-length messages into fixed size blocks. In this paper it is assumed that each message belongs to one of P priority classes. The packets of higher priority classes are sent out from the node m advance of the packets of lower priority classes; packets having the same priority are sent out on a "first-in-first-out" basis. The analysis of the message waiting time in the buffer memory of such a node is presented, assuming that the buffer is equipped with M output channels; i.e., during a single time period at most M packets can be sent out. The results obtained can also be used directly for the analysis of the message path delay in certain s~tore-and-forward commumcation systems, particularly in so-called loop (or ring) communication systems; see [3, 5-9, 12-14]. Recent research has focused on buffers with a single output channel, M = 1 The distribution of the single packet delay is analyzed, for example, in [5, 14], while the distribution of the message delay is analyzed in [3, 6, 7, 9, 12]. However, as will be shown, the results of [3, 7, 9, 12] are biased. Buffers wRh multiple output channels are considered in [10], where the delay distribution of the first packet of the message was calculated in a system wRhout priormes. In [8] the average delay of a single packet was calculated in a system with priorities.
Performance analysis of the single server queue with state-dependent service
International Journal of Systems Science, 1996
In this paper, we investigate a simplified head-of-the-line with priority jumps (HOL-PJ) scheduling discipline. We consider a discrete-time single-server queueing system with two priority queues of infinite capacity and with HOL-PJ priority. We derive expressions for the probability generating function of the system contents and the packet delay. Some performance measures (such as mean, variance and approximate tail distributions) of these quantities are derived and are used to illustrate the impact and significance of a HOL-PJ priority scheduling discipline in an output-queueing switch. We compare this dynamic priority scheduling discipline with a FIFO scheduling and a static priority scheduling (HOL) and we investigate the influence of the different parameters of the simplified HOL-PJ scheduling discipline.
Switching Processes in Queueing Models
Wiley & ISTE, 2008
Data Anisimov, V. V. (Vladimir Vladislavovich) Switching processes in queueing models / Vladimir V. Anisimov. p. cm. Includes bibliographical references and index. ISBN 978-1-84821-045-5 1. Telecommunication--Switching systems--Mathematical models. 2. Telecommunication--Traffic--Mathematical models. 3. Queuing theory. I. Title. TK5102.985.A55 2008 519.8'2--dc22 2008008995
Lecture Notes in Computer Science, 2011
The paper is devoted to the asymptotic investigation of switching queueing systems and networks. The method of analysis uses the limit theorems of averaging principle and diffusion approximation types for the class of "Switching Processes" developed by the author. This class can be used to describe hierarchic stochastic systems with random switching due to internal and external factors.
Sojourn times in a processor sharing queue with service interruptions
2000
We study the sojourn time of customers in an M/M/1 queue with processor sharing service discipline and service interruptions. The lengths of the service interruptions have a general distribution, whereas the periods of service availability are assumed to have an exponential distribution. A branching process approach is shown to lead to a decomposition of the sojourn time into independent contributions, that can be investigated separately. The Laplace-Stieltjes Transform of the distribution of the sojourn time is found through an integral equation. We derive the first two moments of the sojourn time conditioned on the amount of work brought into the system and on the number of customers present upon arrival. We show that the expected sojourn time of a customer that arrives at the system in steady state is not linear in the amount of work he brings with him. Finally, we show that the sojourn time conditioned on the amount of work, scaled by the traffic load, converges in heavy traffic to an exponential distribution. This study was motivated by a need for delay analysis of elastic traffic in modern communication networks. Specifically, the results are of interest for the performance analysis of the Available Bit Rate (ABR) service class in ATM networks, as well as for the best-effort services in IP networks.
The processor-sharing queueing model for time-shared systems with bulk arrivals
Networks, 1971
We consider a model which is applicable to time-muitiplexed systems, such as multiplexed comunication channels and time-shared computing facilities. In this (processor-sharing) queueing model, all jobs currently in the system share equally the processing capability of the server. investigate the processor-sharing model for the case of bulk arrivals. The mean response time of the system as a function of required service time is derived. An example is given to show ths effect of bulk arrivals versus single arrivals for a constant utilization.
Overload Performance of Several Processor Queueing Disciplines for the M/M/1 Queue
1986
In a variety of, overloaded queueing, systems (e+, an overloaded call processing system), long delays can result either in poor service given to the customer or in customers, unknown to the system, turning "bad." For example, in switching systems, long dial tone delays can result in customers initiating,dialing before receiving dial tone. In this case the system will, not receive ali the digits and an unsuccessful call results. This can lead to the system expending real time on unsuccessful services and; therefore, reduces the effective throughput. Thus, there is a deed for control schemes which reillice the load offered to the processor by selectively refusing service to some customers in such a way as to keep delays,, for those customers which are selected for service, small. This fact has been recognized and has led to improved strategies for local switches. In this paper ,we analyze and compare the performance of various queueing and service disciplines for an M / M / l Queue. We consider LIFO and FIFO schemes with customer rejection mechanisms corresponding to pushing,out or timing out older customers in queue. Delay distributions for served customers are obtained and comparisons based upon throughput-delay tradeoffs are presented. For the situation where customers can turn ribad'' at a random time after their arrival, we compare the throughput of good customers. The results presented are a mixture of classical results, which are briefly stated, and new results which are developed in more detail. The numerical results show a dramatic effect of the queueing and service disciplines on the overload performance and a strong dependence of the throughput of successful services on the meclianism for customers turning "bad." Although results are obtained for a single seii-er queue, they can be used to approximately analyze overload control schemes which control access to distributed systems.