This paper brings the idea of MapReduce parallel processing to high energy physics data analysis, proposes a high energy physics data analysis system based on Hadoop framework. It significantly reduces the number of events that need to do further analysis by 2~3 classes by establishing an event TAG information database, which reduces the I/O volume and improves the efficiency of data analysis jobs. It designs proper MapReduce libs that fit for the ROOT framework to do things such as data splitting, event fetching and result merging by using event pre-selection model based on TAG information and MapReduce model of event analysis. A real system is implemented on BESIII experiment, an 8-nodes cluster is used for data analysis system test, the test result shows that the system shortens the data analyzing time by 23% of 4×106 event, and event number of concurrence analysis per second is higher than cluster nodes when adding more worker nodes, which explains that the case analysis cluster has a good scalability.
In order to improve the energy utilization rate in data center on the premise of guaranteeing Quality of Service(QoS), this paper proposes a data center Virtual Machine(VM) adaptive energy-saving mechanism based on the prediction of users’ access quantity, and researches periodicity of users’ visit by Adaptive Holt-Winters(AHW) prediction method. It can adaptively adjust the number of VM according to user visits to improve the utilization rate of VM and achieve the purpose of reducing data center energy consumption. Simulation experimental results show that the Mean Absolute Percentage Error(MAPE) of AHW method is 22.46% and the utilization rate of VM in data center is 97.88% which is promoted by 37.19% compared with the former utilization without using this adaptive energy-saving mechanism, and proves that this energy-saving mechanism has good statistical properties and stability for forecasting the periodic user access, it can be better satisfy the needs of data center energy-efficient.
In the path planning with endpoint directional constraints, traditional A* algorithm is inefficient due to lack of directional guide. To solve this problem, the dynamic guiding A* algorithm is proposed, which introduces dynamic guide points on the basis of the A* algorithm. The dynamic guiding strategies are given, and then a guiding points generation method is proposed. At the same time, a new evaluation function is designed, and the switch method of the guide points is described. Experimental results show that compared with traditional A* algorithm, this new method can find the resultant path which goes from the specific direction of the start point to the fixed direction of the target efficiently as well by setting the proper parameters, while the planning speed can be improved by more than 10 times.
In the new tiered mobile network architecture, this paper proposes a triangle-grid orbit movement and transmission control model. Based on the model, it designs a movement policy of F-connect and proposes a dynamic two-anchor nodes localization algorithm to compute area triangle-grid nodes map. By combining the map and EHCC-communication lemma, it controls the F-collect to collect the data of S. From the simulation, it obviously indicates that this project movement control strategy suits extremely the large-scale Wireless Sensor Networks(WSN). When the network scale achieves certain degree, it can show the superior performance in data collection. The proposed model provides a good solution of the problem on tiered mobile network of instable communication and high-fraction of event loss.
Aiming at the problem of sending redundant in opportunistic routing protocols based on wireless Mesh network coding, this paper proposes a wireless Mesh network coding opportunistic routing protocol based on dynamic redundancy control, and the transmitting process of this class of protocols is modeled and analyzed by linear space theory. It uses null space acknowledgement to evaluate the amount of update information, and dynamically adjusts node sending redundancy according to total update information. In addition, a forwarding priority scheduling mechanism is introduced to avoid duplicate transmissions of the same linear space from different forwarding nodes. Simulation experimental results show that the protocol improves 30%~100% network throughout and 20%~45% normalized transmission cost compared with MORE protocol.
In order to achieve the deterministic transmission of multimedia data in Ethernet for Plant Automation(EPA), this paper analyzes the feasibility of the multimedia data transmit on EPA communication macro cycle and the principle of IP fragment reassembly based on the study of EPA communication scheduling principle. It designs transmission scheduling model of multimedia data in cycle time slice, and a reassemble algorithm of the slice datagram based on linked list insertion method. Aiming at the problem that the most multimedia device can not integrate EPA protocol stack on the market and participate in schedule EPA field layer, it constructs EPA network test platform of accessing multimedia data. Experimental results show that with the case of no disturbance to control data, multimedia data transmission delay based on EPA is 30 ms~36 ms, it can satisfy the acceptable delay-time of 130 ms, delay jitter less than 4 ms, the image is clear and smooth, and the feasibility of multimedia data deterministic transmit on EPA is examined.