دانلود رایگان مقاله کنترل ازدحام در شبکه داده های نامیده شده

Abstract

As a typical Information Centric Networking, Named Data Networking (NDN) has attracted wide research attentions in recent years. NDN evolves today's host-centric network architecture TCP/IP to a data-centric network architecture. It turns the end-to-end connection-oriented transport of TCP/IP into receiver-driven connectionless transport. Compared with the traditional TCP/IP networking, the transport in NDN has new characteristics: Receiver-driven, One-Interest-one-Data, Multi-Source, and Multi-Path. These distinguished features pose new challenges to NDN congestion control mechanisms. This paper presents a comprehensive survey of state-of-art techniques aiming to address these issues, with particular focus on improving the effectiveness of congestion detection and the efficiency of Interest rate shaping. As a new research area, this paper also points out research challenges and open issues in this subject.

5. Conclusion

There are significant differences between the NDN architecture which is centered on content and the traditional TCP/IP architecture which is centered on IP address, and their transport modes are also very different. Therefore, the congestion control mechanisms which are studied and designed in traditional TCP/IP networks cannot be directly adopted in NDN network. The new congestion control mechanisms and transport protocols of NDN network need to be studied. In this paper, we firstly highlight several new features of NDN transport, and point out the congestion control issues in NDN. Then, we analyze the methods of designing congestion control mechanisms for NDN by classifying them into three categories: receiver-based control, hop-by-hop control and hybrid method. The typical existing NDN congestion control algorithms and transport protocols are also classified and analyzed with emphases on their principles and comparison. Based on the deep analysis and discussion in this paper, research challenges and open issues in this field are proposed, which is helpful for researchers to quickly involve in and move forward.