具有两类请求的云计算中心服务器数量的优化

为提高云计算中心的服务质量,节约系统成本,针对具有两类用户请求的云计算中心,提出云计算中心的服务器数量的优化方案。首先,建立了具有两类用户请求的排队模型,分析系统的稳态概率分布、平均队长等性能指标;然后,建立了云计算中心的能耗模型;最后,联合系统的等待成本和能耗成本,构建系统的成本函数,对系统的服务器数量进行优化,从而使系统的成本最小。数值分析结果表明最优服务器数量是用户请求到达率的非减函数,为了使系统成本最小,云计算中心需要动态调整服务器的数量。     

信元调度问题中有优先级的多服务台排队系统

根据排队理论对带有优先级的多服务台信元调度问题进行了定量分析,推导出了每个级别的信元平均时延及丢包率的计算公式,并进行了计算机仿真。根据理论分析和仿真结果,得出了在缓冲区既定的情况下,可以采用提高服务率或者是增加服务台数来有效的降低平均时延和丢包率的结论。由已经得到的公式,可以确定将平均时延及丢包率控制在所允许的范围的服务率和服务台数。     

An optimal deterministic control policy of servers in front and back rooms with a variable number of switching points and switching costs

In this paper we consider a retail service facility with cross-trained workers who can perform operations in both the front room and back room. Workers are brought from the back room to the front room and vice versa depending on the number of customers in the system. A loss of productivity occurs when a worker returns to the back room. Two problems are studied. In the first problem, given the number of workers available, we determine an optimal deterministic switching policy so that the expected number of customers in queue is minimized subject to a constraint ensuring that there is a sufficient workforce to fulfill the functions in the back room. In the second problem, the number of workers needed is minimized subject to an additional constraint requiring that the expected number of customers waiting in queue is bounded above by a given threshold value. Exact solution procedures are developed and illustrative numerical examples are presented. Supported by a grant from the Faculty Research Committee at the C. W. Post Campus of Long Island University     

多服务员时两种等待队列性能的比较

利用排队论的相关知识,对计算机系统中常见的多服务员情况下的排队现象进行分析,通过理论推导、并用实际数据表明了在多服务员模式下,单一共享的排队等待队列的性能要优于多个独自的排队等待队列。     

Optimal assignment of dynamic priorities in queuing systems with two types of customers

We consider queuing systems with two types of customers. For such systems, we develop numerical procedures for computation of optimal dynamic priorities in the case of multiplicative priority functions. The optimality criterion is based on the total queue length for customers of both types. Our technique allows one to take into account waiting-time bounds. The optimization problem is formulated in the language of linear-fractional programming. To illustrate our technique, we present some numerical results at the end of the paper.     

FOPS: A new framework for the optimization with variable number of dimensions

A single‐ and multi‐objective optimization package is presented and described in detail. It contains an ensemble of local and global optimization routines. Procedures controlling variable number of dimensions are implemented as well, which is a rare feature among optimization oriented packages. The package is provided as a MATLAB toolbox. It excels in versatility and extensibility, which is demonstrated on a series of examples covering classical electromagnetism and antenna design. It is taken for granted that defining parameters of the optimization method can be set prior to the simulation run. However, its effective performance can be changed during the optimization run thanks to the full control feature. Moreover, it opens new possibilities in merging various algorithms into hybrids, performing complex dynamic programming tasks, or exploiting third party software. These advantages render the package as a perfect tool to deal with nowadays challenging engineering tasks.     

Optimization analysis for an infinite capacity queueing system with multiple queue-dependent servers: genetic algorithm

This paper considers an M/M/r queueing system with infinite capacity, in which the number of working servers changes depending on the queue length. The steady-state probability distributions and the expected number of customers in the system are derived, which are used to construct a cost function. In order to minimize the expected cost of the system, we use the genetic algorithm to find the best thresholds of queue length in activating servers and their corresponding service rate. Some illustrative examples are provided to demonstrate how the process of this algorithm works for the optimal management policy of the multi-server queueing system.     

Software tools for formation of state spaces of queuing systems with a common type buffer

Universal algorithms and software tools are developed for formation of discrete components of state spaces for queuing systems with a common type buffer. A statement on the parameters of such a buffer is proved. __________ Translated from Kibernetika i Sistemnyi Analiz, No. 3, pp. 164–168, May–June 2007.     

Numerical methods for analysis of multiflow queuing systems with virtual partition of a common buffer

A new approach is proposed to studying queuing systems with request-specific channels and a common finite buffer for requests of different types. For sharing the common buffer, a virtual partitioning strategy is used. Explicit formulas for blocking probabilities of different requests are derived and optimization problems are solved. The results of numerical experiments are presented.Translated from Kibernetika i Sistemnyi Analiz, No. 6, pp. 162–172, November–December 2004.This revised version was published online in April 2005 with a corrected cover date.     

Stationary characteristics of a single-channel queuing system with one waiting space

A semi-Markovian model of a queuing system, with a discrete-continuous phase space of states is developed. Its main stationary characteristics are determined. __________ Translated from Kibernetika i Sistemnyi Analiz, No. 5, pp. 51–62, September–October 2006.     

Ergodicity condition for a system with queuing and servicing of objects with a complicated structure

A single-channel queuing system with a Poisson incoming flow of objects is considered. Each object consists of several spaced requests. A simple ergodicity condition is established. __________ Translated from Kibernetika i Sistemnyi Analiz, No. 5, pp. 8–12, September–October 2007.     

Multi-server retrial model with variable number of active servers

This paper deals with a multi-server retrial queueing model in which the number of active servers depends on the number of customers in the system. To this end, the servers are switched on and off according to a multi-threshold strategy. For a fixed choice of the threshold levels, the stationary distribution and various performance measures of the system are calculated. In the case of equidistant connection levels, the optimum threshold level is numerically computed.     

Modelling priority queuing systems with varying service capacity

Many studies have been conducted to investigate the performance of priority queuing (PQ) systems with constant service capacity. However, due to the time-varying nature of wireless channels in wireless communication networks, the service capacity of queuing systemsmay vary over time. Therefore, it is necessary to investigate the performance of PQ systems in the presence of varying service capacity. In addition, self-similar traffic has been discovered to be a ubiquitous phenomenon in various communication networks, which poses great challenges to performance modelling of scheduling systems due to its fractal-like nature. Therefore, this paper develops a flow-decomposition based approach to performance modelling of PQ systems subject to self-similar traffic and varying service capacity. It specifically proposes an analytical model to investigate queue length distributions of individual traffic flows. The validity and accuracy of the model is demonstrated via extensive simulation experiments.     

Dynamic control of service rates in queuing networks

A method for dynamic control of service rates in closed exponential queuing networks is proposed. The performance of queuing networks with the service-rate control is analyzed, and the main steady-state network characteristics are computed using an analytic approximation. A simple example of a queuing network with controlled service rates is considered as an illustration. The efficiency of the service-rate control is confirmed by Monte Carlo simulations, which, as a by-product, also show acceptable accuracy of our analytical approximations.     

Product from solutions for multiserver centres with concurrent classes of customers

The MultiServer centre with Concurrent Classes of Customers (MSCCC) is a service centre consisting of B parallel identical exponential servers. The customers requesting service at the MSCCC centre belong to K groups. Customers arriving at the MSCCC centre are queued in the order of their arrival. A customer from group k will go into service at the MSCCC centre provided that one or more of the B servers is free and that at most n − 1 other group k customers are in service at the MSCCC centre.

The MSCCC centre can be applied to model systems where customers simultaneously occupy two resources. The system resources are partitioned into K primary and B secondary resources. A customer of group k can access primary resource k if it already is in possession of a secondary resource and if at most n − 1 other group k customers are using primary resource k.

This paper defines the MSCCC centre and presents several examples of computer (sub)systems that can be modelled using the MSCCC centre. The MSCCC centre is shown to satisfy local balance: therefore a multiclass queuing network consisting of BCMP and MSCCC centres has a product form solution. The joint probability distribution (JPD) for a queuing network consisting of several BCMP centres and one MSCCC centre is derived. Aggregation techniques are next used to reduce the JPD to a computationally tractable form. A Mean Value Analysis algorithm is presented for calculating the closed and open chain performance measures at the MSCCC centre.     

基于事件驱动的排队系统仿真器的设计与实现

根据生灭过程的基本原理,从处理排队系统中特定的事件出发,设计并实现了一个基于事件驱动的排队系统仿真器。该仿真器能根据特定的系统配置,对实际数据进行分析和对特定过程进行模拟。仿真器包含的开发接口支持其模拟功能的扩展。实验表明,该仿真器能快速地完成特定的模拟过程,并获得与理论值非常接近的模拟结果,因此能用于对复杂排队问题的仿真和分析。     

Finding a form to hold the act of queuing

In this paper, we present an initial appliance idea that is to reflect the practice of queuing. This is to enable exploration on how the act of queuing might evolve with new appliances and experiences. Based on the notion that a good appliance is one with limited functionality, we have worked on a limited number of excerpts from the act of queuing. By using interactional features from the field as material for design, these features are formed into the Queuing Display, which reflects issues on queuing as such.     

超市排队系统仿真研究

针对超市顾客等候过久问题,基于Agent范型,建立了顾客排队系统仿真模型。将顾客、收银台抽象为不同的Agent,并引入收银台管理Agent和引导员Agent。通过不同Agent之间的交互仿真超市排队系统的运行。模型中给出了一种收银台快速结算动态调整算法。实验数据和分析显示,该模型不仅可以真实地模拟超市排队系统运行现象,而且减少了顾客平均等待时间,降低了超市运营成本。     

Optimizing utilization of resource pools in web application servers

Among the web application server resources, the most critical for their performance are those that are held exclusively by a service request for the duration of its execution (or some significant part of it). Such exclusively held server resources become performance bottleneck points, with failures to obtain such a resource constituting a major portion of request rejections under server overload conditions. In this paper, we propose a methodology that computes the optimal pool sizes for two such critical resources: web server threads and database connections. Our methodology uses information about incoming request flow and about fine‐grained server resource utilization by service requests of different types, obtained through offline and online request profiling. In our methodology, we advocate (and show its benefits) the use of a database connection pooling mechanism that caches database connections for the duration of a service request execution (so‐called request‐wide database connection caching). We evaluate our methodology by testing it on the TPC‐W web application. Our method is able to accurately compute the optimal number of server threads and database connections, and the value of sustainable request throughput computed by the method always lies within a 5% margin of the actual value determined experimentally. Copyright © 2010 John Wiley & Sons, Ltd.