论文一
Time: 1.7 hours
Qi Yu, Manjeet Rege. A Relational approach for efficient service selection. ICWS(Application and Industry Session#Services Optimization), 2009
作者Qi Yu, 看了一下简历, 原来是校友, 我到浙大的这一年他正好本科毕业, 目前是他是Rochester Institute of Technology的Assistant Professor.
发表的相关论文有:
Framework for Service Query Algebra and Optimization, ACM TWEB, 2008
Deploying and Managing Web Services: Issues, Solutions, and Directions, VLDB Journal, 2008
以下是论文摘录:
1. QoWS(Quality of Web Service)
作者自己提的一个缩写, 没什么特别的含义, 就是指web service的quality属性, 文中涉及的有Latency, Reliability, Availability, Fee, Reputation
2. 本文的贡献有
(1) 提出了一种service selection approach, 能够处理"局部匹配的情况"
(2) the service selection approach is built upon the service query optimization framework
在这个framework中, 定义了service model(相当于abstract class).
提出了两种分解service relation的方法: Q-Decompose, O-Decompose
支持使用标准的关系数据库操作来进行服务选择.
3. 建模方法(S2.1)
Service Schema (SC) 
表中有三个web service, 每个service包含数个operation, 每个service对应一个DAG图. DAG图表示service内各个操作调用的依赖关系;此外还存在inter-service dependency(图中虚线).
Operation Graph (OG) 
对于给定的SC, OG是一种映射关系:
an operation op --> the union of all the paths in SC that lead to op
4. 看了abstract和introduction后,还是不太清楚. Abstract里介绍本文工作的第一句话: "We present in this paper a systematic approach for efficiently service selection by using QoWS as the major criterion.", Introdution里介绍本文工作的第一句话: "We present in this paper a service selection approach that can deal with providers that partially conform to the abstract services." 感觉这两句话里说的是不同的两件事情, 难免让读者有些confusing. 没有通读全文.
论文二
Time: 1.2 hours
David Chiu, Sagar Deshpande, Gagan Agrawal, Rongxing Li. A Dynamic Approach toward QoS-Aware Service Workflow Composition, ICWS(Application and Industry Session 7#Services Composition), 2009
作者David Chiu是OSU的博士生(预定明年毕业), 博士论文的题目是"Auspice: Automatic Service Planning in Cloud/Grid Environments". 关于OSU, 校内BBS上有人戏称是"浙大北美分校", 言下之意浙大去OSU的人多(一方面是因为浙大学生申请的多, 另一方面是因为OSU喜欢招浙大的学生).
以下是论文摘录:
1. 背景: "Such systems often support user preferences, e.g., time of completion, but with the rebirth of distributed computing via the grid/cloud, new challenges are abound: multiple disparate data sources, networks, nodes, and potential for moving very large datasets."
2. 这篇文章主要是在介绍作者的系统"Workflow Broker". 这个系统类似service composer, 接受用户的请求, 自动生成组合计划并执行.
该系统分为4个模块: Query Decomposition Layer, Semantics Layer, Planning Layer, Execution Layer.
本文着重讨论Planning Layer, 这个Layer的功能 "for handling QoS constraints while composing service workflows".
3. 本文对服务组合的建模采用了recursive的方式: 
ε表示一个空流程, d是data instance, (op, Pop)中op是service operation, Pop = (p1, …, pk), 其中pi又是一个workflow.
Pop里的流程似乎是指w的后续流程
流程涉及两种cost:
(1) time cost: 包括service execution time和transmission time
(2) error cost: "error estimation of a given workflow"
4. 本文讨论的workflow侧重scientific domains.
5. 没有通读全文, 不是很感兴趣, 摘录一下留做纪念.