Metadata management for quality-driven delivery

分布式系统和无线技术的最新进展，以及复杂应用程序的开发，现在允许用户从任何地方使用任何设备访问任何信息。在这种复杂环境的背景下，更具体地说，对于连续媒体递送，引入了服务质量(Qos)管理来支持具有质量要求的服务提供，传统上关注于变化的通信条件。如果我们从更广泛的角度考虑服务质量的概念，我们必须将用户放在流程的中心，并允许他表达根据高级别质量维度指定的非功能需求，这些维度随后被映射到低级别维度。为此，我们提出了一种质量驱动交付(Quality-Driven Delivery，QDD)方法，旨在通过集成质量信息管理、系统组件互操作、QDD活动的分布式执行或多准则优化等不同机制来提供面向用户的QDD管理。在我们的研究计划框架内，我们将重点关注质量信息管理，以支持分布式多媒体应用的QDD。质量信息可以是描述计算或通信资源、要交付的数据、技术基础设施或用户特征的任何类型的元数据。这类信息对于支持QDD活动和决策至关重要，应提供存储、访问、共享、转移、整合或交换高质量信息的机制。我们将研究模型管理和模型工程方法，以提出、设计和实施此类机制。模型管理和模型工程最近在数据库、软件工程和语义网研究界引起了极大的兴趣。我们认为，它们特别适用于设计和制定灵活和可扩展的元数据管理战略。我们计划通过在媒体艺术、医疗保健和视频传输领域的具体应用来验证我们的方法。

The recent advances in distributed systems and wireless technology, as well as the development of sophisticated applications, now allow users to access any information, from anywhere with any device. In the context of such complex environments, and more specifically for continuous media delivery, Quality of Service (QoS) management has been introduced to support service provision with quality requirements, traditionally focusing on varying communication conditions. If we consider the notion of QoS from a broader perspective, we have to position the user at the center of the process and allow him to express non-functional requirements specified according to high-level quality dimensions that are later mapped onto low-level dimensions. For that purpose, we propose a quality-driven delivery (QDD) approach, aiming at providing user oriented QoS management through the integration of different mechanisms, such as quality information management, system components interoperability, distributed execution of QDD activities or multi-criteria optimization.   In the framework of our research program, we will focus on quality information management to support QDD for distributed multimedia applications. Quality information can be any type of metadata describing the computing or communication resources, the data to be delivered, the technical infrastructure or the user characteristics. Such information is essential to support QDD activities and decisions, and mechanisms should be provided to store, access, share, transfer, integrate or exchange quality information. We will investigate model management and model engineering approaches to propose, design and implement such mechanisms. Model management and model engineering have recently attracted a lot of interest in database, software engineering and semantic web research communities. We believe that they are particularly adapted for the design and development of flexible and extensible metadata management strategies. We plan to validate our approach with specific applications in the fields of media arts, health-care and video delivery.