Sleep Time Decision by First and Last Packet Arrival Time After REPORT in Ethernet Passive Optical Networks with Sleep Mode and Delay Constraint

  • 許 恩榮

Student thesis: Master's Thesis

Abstract

Abstract– Concerns about energy efficiency in the Ethernet passive optical network (EPON) are growing In response enabling optical network units (ONUs) to enter low-power sleep mode has become a focus in reducing energy consumption However in sleep mode packets can not be transmitted and so packet delays increase Accordingly the sleep time decision of when the ONU can enter sleep mode to conserve energy yet not violate the delay constraints becomes a challenge Green Bandwidth Allocation (GBA) was used to study the analytical model and propose a closed-form expression to determine the sleep time In this manner sleep time can be maximized for energy savings while simultaneously ensuring the average packet delay can satisfy the delay constraint However the packet arrival in the network may not be the same as the analytical assumption In this thesis we start from determining the sleep time from a real-time information: first packet arrival time after REPORT message to keep packets satisfying delay constraint and meanwhile conserve energy as great as possible Nevertheless the sleep time decided by first packet arrival time causes a frequent transition overhead which wasted a lot of energy Hence the last packet arrival time after REPORT message and a ratio parameter is added to enlarge sleep time so as to decrease transition overhead frequency Since the OLT lacks packet arrival time information the extension to the REPORT message carrying additional first and last packet arrival times after the last REPORT is sent to the OLT Moreover since the OLT needs to respond to the real-time arrival information an on-line separated polling scheme and scheduling is presented Simulation results showed that the percentage of packets violating the delay constraint can be lower than the GBA; however the energy saving difference is no more than 1%
Date of Award2015 Aug 26
Original languageEnglish
SupervisorChuan-Ching Sue (Supervisor)

Cite this

'