A Power Saving Scheduler with Packet Concatenation for D2D Communications

• Main Authors: I-Feng Chen, 陳奕峯
• Other Authors: Zse-Hong Tsai
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/96415154625870831668

碩士 === 國立臺灣大學 === 電信工程學研究所 === 101 === Internet of things(IoT) has been predicted to grow up in wired and wireless communication and become the most popular industry in ten years. In the scenario of IoT, many devices, called D2D devices, will connect each other by wired or wireless network. It is not suitable to connect by wired network in some environment, so connecting through wireless network is familiar. However, many kind of application working on devices need a powerful data center to manage and process the information which is obtained by the D2D devices. When information has been received at the data center, professionals will analysis, process, and make some decision by the information. When huge data produced by D2D devices transmits through wireless network and core network to data center, it becomes a big issue for ISP that how to reduce power consumption and lower packet overhead. However, current network interface cards (NIC) equipped by devices of mobile communication network are with low utilization, but these devices are continuously consuming power. The most fundamental method to lower power consumption is turning the equipment off, but it may causes that some packets can`t be routed through the equipment and many users have bad user experience. On the other hand, it is obvious that devices will consume higher power when transmitting data. According to research, devices will consume additional power when transmitting a piece of data, it is called tail energy. In this thesis, we present packet concatenation algorithm to reduce tail energy and it could provide good QoS. In this thesis, we present channel capacity allocation algorithm, network selection algorithm, and packet concatenation algorithm to reduce power consumption of network equipments where full of IoT. At the same time, this algorithm can also reduce packet overhead, so that each network equipment can support higher data rate. Simulation results illustrate that power consumption when transmitting 1 MB can lower 46.77% of original power under high traffic load, and power consumption can reduce 57.56% of original power if under light traffic load. Inaddition, packet overhead is reduced to 5.5% from 29.28%.