移动协同工作中基于AST的一致性维护技术研究

The rapid development of new mobile applications (e.g. mobile microblogging, mobile online community, etc.) provides a real-time interactive environment for people, and gradually meets people's collaborative needs. Small storage space, limited power, small screen size of mobile devices as well as the instability of mobile environments pose a challenge to consistency maintenance techniques in traditional cooperative work. This project will take full consideration to the characteristics of mobile environments, and conduct researches on Address Space Transformation (AST) consistency maintenance approaches under the premise of saving storage space and supporting large-scale dynamic user cooperative needs. Weak consistency model is created, and by changing the granularity of consistency maintenance, different needs of the collaboration are met; Document localization model is created, sliding window and dynamic segmentation data strategy is presented to avoid the storage space problem with full replication strategy; Considering equipment frequently dropped in the mobile environment, sequence sort and merge algorithm and string operation merge and convert algorithm is proposed; Timestamp-free AST algorithm is designed to support dynamic large-scale users' join and exit. A Prototype system based on the above study is presented and after that, user behavior analysis is conducted.

手机微博、移动在线社区等新型移动应用的迅速发展为人们提供了一个可以实时交互的环境并逐渐满足人们的协同需求。移动设备存储空间小、电量有限、屏幕尺寸小等设备自身特点和移动环境网络状态不稳定的问题对传统协同工作中的一致性维护技术提出了挑战。本课题将充分考虑到移动环境的特点，在节省存储空间、支持大规模动态用户协同需求的前提下研究基于地址空间转换（AST）的一致性维护方法。创建支持弱一致性维护的模型，通过改变一致性维护的粒度满足不同的协同需求；建立文档局部化模型，通过滑动窗口和动态分割的数据副本策略避免全复制策略带来的空间存储问题；针对移动环境中设备频繁掉线的问题提出序列排序合并算法和串操作的合并转化算法；设计不依赖于向量时间戳的AST算法支持移动环境动态大规模用户的加入与退出。基于上述研究设计实现原型系统，并进行用户行为分析。

本项目组针对移动设备存储空间小、电量有限、屏幕尺寸小等设备自身特点和移动环境网络状态不稳定的问题，研究移动协同工作中基于AST的一致性维护方法，以满足移动环境下动态大规模用户实时协同工作的需求。主要创新成果包括：移动和异构云环境下基于AST的模型构建；移动环境下的数据局部复制策略；支持混合粒度操作的基于标量时间戳的新型AST算法；在线协同编辑社区中基于内容质量的推荐技术；在线社交网络中基于用户隐私保护和用户影响力的推荐技术；基于AST的移动协同应用原型系统实现及其用户行为分析研究。这些成果发表在高水平国际会议和期刊上，其中CCF A类会议论文7篇，SCI期刊论文2篇。在本基金项目资助下，项目组共发表学术论文20篇，其中SCI收录2篇，EI收录18篇；申请国家发明专利3项；出站博士后1名，培养博士8名、硕士16名；组织国内学术研讨会3次，参加国际学术会议24人次。