| Due to requirement of QoS by high-speed networks, software can't deal packets fast enough. More and more researchers are dedicated themselves in ASIC or FPGA design. This paper provides research basis of some mainly scheduling algorithms. And according to simulation experiments of some algorithms, we may find some factors, such as total session numbers, size of memory, scheduling strategy and character of traffic, that affect implementation of scheduling algorithms. The main purpose of this paper is to provide the basis of theory researching in some scheduling algorithms.A rough arrangement of this paper is as follows. In chapter two, fundamental knowledge about networks related to scheduling is discussed. Some QoS parameters are discussed in ?.1. Fundamental queuing theories, including model M/M/1 and Jackson queuing network, are discussed in ?.2. In chapter three, disciplines and strategies of scheduling algorithms are discussed. In ?.1 and ?.2, basic scheduling knowledge is provided. Work-conserving and non-work-conserving and function of shapers are respectively discussed in ?.3 and ?.4.In chapter four, Packet Fair Queuing (PFQ) algorithms are discussed. After discussing GPS, the origin of PFQ algorithms, some PFQ algorithms such as shVC, WFQ and WF2Q will be discussed in ?.3. And in chapter five, Round Bobin (RR) and Service Curve-based algorithms are discussed. WRR algorithms are discussed as examples of RR algorithms in ?.1. At the same time, some simulation experiments of an improved WRR algorithm are given in ?.2. And in ?.3, HFSC algorithm is provided as an example in Service Curve-based algorithms.Then, thorough explanations to how to implement scheduling algorithms in hardware can bs found in chapter six. And some practicable solutions to diminish comparison of timestamps are provided for further understanding.At the end of every previous chapter, there is a brief summary. And in last chapter seven, I give a summary of the all algorithms previous discussed. |
PDF Full Text Request