QoS and QoE

1.林闯 胡杰 孔祥震，用户体验质量(QoE)的模型与评价方法综述，计算机学报，2012 January、Vol35、Num1.

　　http://cjc.ict.ac.cn/qwjs/view.asp?id=3529

　　由于用户的主观因素难以测量和量化，给QoE的建模和评价带来了难点。论文综述了QoE的影响因素、量化方法和评价方法，涉及到一些统计、心理方面的知识，最后给出了短信服务QoE评价体系。QoE算QoS升级版，更加注重以人为本，IEEE Network做过相关专辑（IEEE special issue on QoE- March 2010 -地址见http://dl.comsoc.org/livepubs/ni/public/2010/mar/index.html）。

2.Jorge Cardoso1, Amit Sheth2, John Miller2, Jonathan Arnold3, and Krys Kochut2,Quality of service for workflows and web service processes

　　对工作流QoS的建模、预测、计算，考虑了time、cost、reliability，对于Task t

　　T(t) = DT(t) + PT(t)，Delay time (DT)表示发出请求到开始处理间的时间，如排队与启动时间；Process time (PT)处理时间；

　　C(t) = EC(t) + RC(t)，enactment cost (EC)工作流系统管理监督运维（direct labor, machine, direct material, and setup带来的花费，realization cost (RC)执行过程带来的花费；

　　R(t) = 1 – (system failure rate + process failure rate)，System failure rate为（unsuccessful executions)/(called for execution），process failure rate为failed or aborted)/((failed or aborted) + (done or commit)).

　　论文比较古老，洋洋洒洒40余页，内容比较细，写introduce和综述可以参考。

3.Duhang Zhong, Zhichang Qi,Xishan Xu，Reliability Prediction and Sensitivity Analysis of Web Services Composition， 

论文中对可靠性的定义为：the probability of its ability to successfully carry out its own task when it is invoked。

4.Vuong Xuan Tran a, Hidekazu Tsuji b, Ryosuke Masuda，A new QoS ontology and its QoS-based ranking algorithm for Web service，Simulation Modelling Practice and Theory 17 (2009)

　　由于是期刊论文，所以篇幅较长，总共21页，前面几节对QoS建模进行了详细的综述，讲述了QoS模型的基本功能、属性、要求。

 

 

　　MTBF——全称是Mean Time Between Failure，即平均无故障工作时间。就是从新的产品在规定的工作环境条件下开始工作到出现第一个故障的时间的平均值。MTBF越长表示可靠性越高正确工作能力越强 。 Reliability

　　MTTR——全称是Mean Time To Repair，即平均修复时间。是指可修复产品的平均修复时间，就是从出现故障到修复中间的这段时间。MTTR越短表示易恢复性越好。Availiability 

　　MTTF——全称是Mean Time To Failure，即平均失效时间。系统平均能够正常运行多长时间，才发生一次故障。系统的可靠性越高，平均无故障时间越长。

5.EM Maximilien, MP Singh，A framework and ontology for dynamic web services selection， Internet Computing, IEEE, 2004

　　Availability is the probability that a service can respond to consumer requests. It has two subclasses: MTTR (mean time to repair, meaning the average time for restoring a failed service) and UpTime (the duration for which the service has been operational continuously without failure). Availability is mildly parallel to reliability and typically mildly opposite to capacity。

　　Capacity is the limit on the number of requests a service can handle. When a service is operated beyond its capacity, its availability and reliability are negatively affected。

　　Economic captures the economic conditions of using the service. Usage cost is a key economic attribute。

　　Performance characterizes performance from the consumer’s perspective. Examples are Throughput (the rate of successful servicerequest completion) and ResponseTime (the delay from the request to getting a response from the service).

　　Reliability  is the likelihood of successfully using a service. Typically, it parallels availability, but its main aspects also include Fault Rate (the rate of invocation failure for the service’s methods); MTBF (mean time between failures); Consistency (the failure rate’s lack of variability); Recoverability (how well the service recovers from failures);  Failover
(whether the service employs failover resources, and how quickly); and  Disaster resilience (how well the service resists natural and human-made disasters).

　　Robustness is resilience to ill-formed input and incorrect invocation sequences.

　　Scalability defines whether the service capacity can increase as needed.

　　Security captures the level and kind of security a service provides. Its key components include Auditability (the service maintains auditable logs); Authentication (the service either requires user authentication or accepts anonymous users); Encryption (the type and strength of encryption technology used for storage and messaging); and NonRepudiation (whether consumers can deny having used the service)