An Efficient Bandwidth Reservation with Modified Two-Pass RSVP Setup Mechanism

碩士 === 國立中山大學 === 電機工程學系 === 87 === In this thesis, we propose an efficient bandwidth reservation with modified two-pass RSVP setup mechanism. Existing RSVP setup mechanism (OPWA) has killer reservation problems when merging RSVP reservation requests. There were several previou...

Full description

Bibliographic Details
Main Authors: Guan-Ying Pao, 包冠英
Other Authors: Tsang-Ling Sheu
Format: Others
Language:zh-TW
Published: 1999
Online Access:http://ndltd.ncl.edu.tw/handle/10855601189572808874
Description
Summary:碩士 === 國立中山大學 === 電機工程學系 === 87 === In this thesis, we propose an efficient bandwidth reservation with modified two-pass RSVP setup mechanism. Existing RSVP setup mechanism (OPWA) has killer reservation problems when merging RSVP reservation requests. There were several previous works being proposed to resolve the killer reservation problems. However, these solutions are very complicated and may waste bandwidth due to excessive reservation. Therefore, we propose a modified two-pass with advertising (MTPWA) mechanism to efficiently avoid the killer reservation problems. This mechanism utilizes both the advantages of one-pass and two-pass setup mechanisms to improve the drawbacks of using pure two-pass. MTPWA differs from previous two-pass setup mechanisms in that it will request pre-engaged bandwidth at the first-pass before actually reserving bandwidth at the second pass. The pre-engaged bandwidth is a tagged free bandwidth used to deliver non-QoS traffic. Since the bandwidth can be dynamically reserved at the first pass by using bounded flowspec and greedy policy, the killer reservation problems are avoided. At the second pass, there will be sufficient bandwidth for the receivers to use by converting pre-engaged bandwidth to reserved bandwidth. We analyze the characteristics of pre-engaged bandwidth in this thesis first, and then we build a mathematical model based on Markov Chain to evaluate the efficiency of the proposed MTPWA. Finally we use programs to simulate the MTPWA mechanism and verify our mathematical results.