Multi-tier service differentiation by coordinated learning-based resource provisioning and admission control

Multiple Internet applications are often hosted in one datacenter, sharing underlying virtualized server resources. It is important to provide differentiated treatment to co-hosted applications and to improve overall system performance by efficient use of shared resources. Challenges arise due to multi-tier service architecture, virtualized server infrastructure, and highly dynamic and bursty workloads. We propose a coordinated admission control and adaptive resource provisioning approach for multi-tier service differentiation and performance improvement in a shared virtualized platform. We develop new model-independent reinforcement learning based techniques for virtual machine (VM) auto-configuration and session based admission control. Adaptive VM auto-configuration provides proportional service differentiation between co-located applications and improves application response time simultaneously. Admission control improves session throughput of the applications and minimizes resource wastage due to aborted sessions. A shared reward actualizes coordination between the two learning modules. For system agility and scalability, we integrate the reinforcement learning approach with cascade neural networks. We have implemented the integrated approach in a virtualized blade server system hosting RUBiS benchmark applications. Experimental results demonstrate that the new approach meets differentiation targets accurately and achieves performance improvement of applications at the same time. It reacts to dynamic and bursty workloads in an agile and scalable manner.     

EFFECT OF MAGNETIZATION OF WATER ON INDUCTION TIME AND GROWTH PERIOD OF NATURAL GAS HYDRATE

The effect of diluted solution's magnetization on induction time and growth period of natural gas hydrate (NGH) has been investigated in quiescent reaction system at pressure of 4.5 MPa and temperature of 274 K with SDS as surfactant, by using volume fixed and pressure falling method. Experimental results show that magnetization will have effect on the induction time of NGH. After magnetization with magnetic field intensity of 0.33 T, the induction time of NGH has been reduced to 47 min (average) from 99 min (average) in which there is no magnetization. On the other hand, the induction time has been prolonged after magnetization of the diluted solution with magnetic field intensity of 0.05 T, 0.11 T, 0.22 T, 0.44 T. Especially with magnetic field intensity of 0.11 T, the induction time had even been prolonged to 431 min (average). The effect of magnetization on the growth period of NGH has not been found at the experimental condition.     

氰化亚铜生产新工艺研究

在氯化钠浓溶液中通氯气将铜氧化生成NaCuCl2，再加入氰化钠溶液生成氰化亚铜沉淀。沉淀经氯化钠浓溶液洗涤和水洗干燥后即得产品。工艺生产周期短，设备简单，成本低，产品质量高，无废气，废渣和有毒废水排放。     

新型甲醇汽油的制备及性能测试

本文研究了甲醇汽油体系相稳定剂,并利用90~#基础汽油和工业精甲醇为主要原料,添加相稳定剂、防腐剂、消烟剂、蒸汽压调节剂和燃烧催化剂等,经科学配方和工艺,成功制备出高性能符合97~#无铅车用汽油国标的甲醇汽油产品。     

具有负载分享的P2P IPTV重迭网络的设计

针对在IPTV服务中当网上收视人数大规模增长时网络带宽消耗及服务器负载大大提高的问题,提出了利用P2P(Peer-to-Peer)概念中的应用层群播所形成的重迭网络来解决IPTV负载分享的问题,并通过P2P中Chord路由算法,加速构建与寻找对应的群播代理人节点。实验证明将Chord与应用层群播代理人做有效结合,可以提供用户具有服务质量、扩充性与负载分享的全方位P2P IPTV服务。     

基于BACnet/IP的设备与实时数据库通信实现

针对不同厂商、不同系统的楼宇自控设备互操作、互通信的局限性,分析了BACnet/IP技术和实时数据库,设计开发了一种基于BACnet/IP技术的实时数据库DasRdb的设备驱动程序。该设备驱动程序采用UDP通信方式,通过调用每个BACnet/IP设备的标准接口函数来驱动实时数据库DasRdb分别进行写操作和读操作,能够有效地实现BACnet/IP设备与实时数据库DasRdb之间的通信,从而使楼控设备的互操作成为可能。现场工程应用表明,该程序的运行契合BACnet的一致性要求,具有简单、高效、精确度高、实时性好的特点。     

时间Petri网与CAGA算法相结合的并行测试研究

在并行测试的研究中对于任务的测试次序一直是个研究重点及难题。用时间Petri网建立关于并行测试系统的模型,并首次用元胞遗传算法(CAGA)对模型中的变迁序列进行搜索,快速得到最好的测试序列。通过仿真结果表明,该算法展示了很好的收敛效果,并且与已有的研究成果相比更具优势。     

基于四元数和增广矩阵束的MIMO雷达角度估计算法

通过研究多输入多输出（Multiple input and multiple out, MIMO）雷达的角度估计算法,基于收发共址的十字阵MIMO雷达系统,将四元数理论应用到 MIMO雷达角度中,提出了一种新的参数估计算法。通过构造四元数模型,结合增广矩阵束（ Matrix enhancement and matrix pencil, MEMP）方法构造增广矩阵,并证明该矩阵的秩 等于目标总数,且不受目标相干性的影响,结合ESPRIT算法实现了对MIMO相干目标的角度估 计。算法无需谱峰搜索,能够实现参数的自动配对,同时降低了运算复杂度。仿真实验进一 步验证了算法的有效性。     

玻璃纤维的形态对氯氧镁水泥增韧效果的影响

本文对短切、连续布置的玻璃纤维对氯氧镁水泥抗弯强度、抗冲击强度、韧性指数的影响进行了研究.结果表明:短切纤维增强氯氧镁水泥基体的增韧效果比连续纤维更好.     

铑含量对Pt-Rh-Pd/Al2O3催化剂的氧化还原性能的影响

应用微反应装置研究了铑含量对Pt-Rh-Pd/Al2O3催化剂的催化活性以及吸附性能的影响.结果表明Rh含量在0%～0.015%之间时,催化剂对CO的活性曲线呈波浪形,对NO的活性曲线为增函数;当Rh含量超过0.015%时,增加Rh的量对CO和NO的转化率几乎无影响.在无助的情况下,用Al2O3为载体制作的Pt-Rh-Pd/Al2O3催化剂,其Pt、Rh、Pd之间的最佳比例为15320.Rh含量的增加有利于氧化性气体的吸附,不利于还原性气体的吸附.