随着计算机网络和通信技术的迅猛发展，数字家庭网络成了通信领域的新热点，也是“3C融合”的关键所在。如何实现数字家庭网络中广泛存在的各种信息设备的互联互通互操作是目前数字家庭发展急需解决的重要问题。由联想提出的闪联（IGRS 信息设备资源共享协同服务）协议正是为了实现企业、家庭等网络中的信息设备间的互联互通互操作而制定的一个简捷高效、易于操作的标准协议。 本文根据作者在实习期间参与的具体项目，对闪联协议做了较深入的研究分析。该协议是一个应用层的网络协议，主要应用于无线音视频的传输。协议定义了如何在有限范围网络域内，通过有线或无线的互连方式，实现多种信息设备、家用电器、通信设备之间的设备自动发现、动态组网、资源共享和协同服务。 从协议栈分模块开发的整体架构出发，设计并实现了协议栈中设备组模块的各个功能函数的流程图和代码。 同时针对设备组模块中存在的有关设备分组和选举的问题，提出了基于目标函数的模糊ISODATA聚类算法和基于模糊等价矩阵的聚类算法，分别解决了设备分组和设备选举的问题。并通过具体的例子分析证明，使用这两种算法达到的效果与人们的在现实生活中的预期是基本一致的。这些工作使协议中的设备分组和选举功能具有了更高的智能性和易用性。最后将这两个模糊聚类算法模块化，并设计了在协议栈系统中实现这两个算法模块的方法和流程。采用这两个算法的闪联协议系统能够自动合理的完成设备分组和选举。

With the swift development of computer network and communication technology, Digital Home network has become a new focus of the communication field and is also the key of 3C fusion. How to realize the interconnection and interoperation across various information devices existing in the Digital Home network is the main problem, which needs to be resolved in the Digital Home development. IGRS (Intelligent Grouping and Resource Sharing) protocol was put forward by Lenovo to provide capacity of interconnection and interoperation across various information devices existing in the corporation or home network. It is a simple, high-efficient and easy technology. According to the project details which the writer participated in the project of the practice, this paper makes an in-depth analysis of the IGRS. The protocol is an application layer protocol and mainly used to transmit wireless audio and video information. It defines how to realize device auto-discovering, dynamic inter-networking, sharing resource and collaborating service among many kinds of information devices in the wire or wireless LAN. By that the protocol was designed and realized with separating module method, every function flow chart and code of the device group module is designed and realized in the IGRS and two new algorithms are proposed to finished unsolved problems of device group module. One is based on target-function fuzzy ISODATA clustering algorithm to solve the problem of device grouping, and another is based on fuzzy equivalent matrix clustering algorithm to solve the problem of device selection. Through the analyzing of practiced examples, using the two algorithms makes device grouping and selection more intelligent and easier, and the effect of them is accordant to the anticipated desire. At last the two algorithms are modularized, and the method and flow charts of realizing of them are designed, which can let system reasonably and automatically finish the function of device grouping and selection.