廉价冗余磁盘阵列（RAID）Cache浅析

廉介冗余磁盘陈列技术已掀起研究开发热潮，磁盘Ｃａｃｈｅ技术的研究早在七十年代就已广泛展开，但是关于磁盘阵列Ｃａｃｈｅ技术的专门性研究文献在国内外并不多见。本文论述了磁盘阵列中引入高速缓存Ｃａｃｈｅ的必要性，综述了磁盘阵列Ｃａｃｈｅ技术的国内外技术动态，提出了磁盘阵列Ｃａｃｈｅ研究中的几个关键问题，并阐述了作者的观点。     

磁盘阵列的应用前景及发展趋势

将对目前世界上广泛关注的磁盘阵列的应用前景作一详细分析和论述，主要包括磁盘阵列产品在各个领域中的应用前景及磁盘阵列技术的发展趋势。其目的在于引起国内对磁盘阵列产吕研究、开发的高度关注。     

磁盘阵列与双机容错在图书馆网络系统中的应用

介绍了磁盘阵列和双机客错技术，说明了磁盘阵列和双机容错在图书馆网络系统中的应用，提出了几点建议。     

磁盘阵列技术及其发展趋势

本文详细介绍了磁盘阵列技术的产生及其发展，重点介绍了目前较为流行的廉价磁盘冗余阵列ＲＡＩＤ技术，通过对磁盘阵列技术的介绍，探讨了今后一些有发展前途的新型存储技术。     

改善磁盘阵列性能的方法

磁盘阵列是解决计算机Ｉ／Ｏ瓶颈问题的有效方法之一，通过对现有磁盘阵列结构的研究，提出了４种改善磁盘阵列性能的方法，即良分利用磁盘带宽，平衡多盘的负载，减少奇偶检验数据存取时间和磁盘阵列Ｃａｃｈｅ技术等，分析结果表明：在事务处理应用领域，利用Ｃａｃｈｅ来把小写转化为大写可大大改善目前一般情况下，以一道数据大小来作为磁盘的基本存取大小是合适的选择，磁盘阵列负载平衡设计不足是指正常模式下系统请求在多盘间     

冗余交叉磁盘阵列性能研究

本文建立了冗余交叉磁盘阵列的队列模型，单盘和同步交叉磁盘阵列采用Ｍ／Ｇ／１排队模型，异步交叉磁盘阵列采用Ｆｏｒｋ／Ｊｏｉｎ排队模型，并利用这些模型对冗余交叉磁盘阵列的性能进行了分析，结果表明在系统请求率较大时，同步交叉磁盘阵列的性能大大优于异步交叉磁盘阵列，而系统请求率较小时，异步交叉磁盘阵列的性能与同步交叉磁盘阵列的性能差别不大，但随系统请求率和数据块的增大差别越来越大。     

小型高可靠磁盘阵列技术若干问题的研究

高可靠磁盘阵列是采用相对廉价的、小容量的高性能磁盘驱动器为单元，通过一定的方式，组成磁盘阵列以提高磁盘容量，提高数据的传输率，目前这一技术已经在国内外引起计算机界的广泛关注。本文将对廉价磁盘冗余阵列体系结构、磁盘阵列控制器及实现过程中理论上和工程上需要解决的若干问题进行一些探讨和研究。     

提高磁盘阵列小写性能的方法研究

磁盘阵列技术对于小写请求不能很好地发挥其优越性。在分析小写问题的成因后，研究了提高磁盘阵列小写性能的几种典型方法，并在此基础上提出了一种针对小写问题的新型体系结构。     

磁盘阵列技术及其选择

存储技术一直以来是计算机技术的热点问题.为了满足用户对存储数据的安全、存取速度快和超大存储容量的需求,磁盘阵列技术得到了迅速发展.通过介绍磁盘阵列技术概念和级别,进一步对不同级别的磁盘阵列性能进行比较,以帮助用户选择合适的磁盘阵列产品.     

不断发展的磁盘阵列技术

介绍了存储系统面临的挑战及I/O系统的瓶颈问题 ,讨论和比较了不同的磁盘阵列技术 ,重点论述了正交结构和Crosshatch结构磁盘阵列技术 ,最后对磁盘阵列技术的现状及前景进行了介绍。     

Algorithm for converting a forest of quadtrees to a binary array

An algorithm for reconstructing a binary array of size N sx N from its forest of quadtree representation is presented. The algorithm traverses each tree of the forest in preorder and maps each ‘black’ node into the spatial domain. The time complexity in mapping is O(log N × B_n + B_p), where B_n is the number of black nodes in the forest and B_p is the number of black pixels in the N × N array. The algorithm has been implemented on an Apple II.     

Impedancemetric gas sensor based on Pt and WO3 co-loaded TiO2 and ZrO2 as total NOx sensing materials

Pt-loaded metal oxides [WO₃/ZrO₂, MO_x/TiO₂ (MO_x = WO₃, MoO₃, V₂O₅), WO₃ and TiO₂] equipped with interdigital Au electrodes have been tested as a NO_x (NO and NO₂) gas sensor at 500 °C. The impedance value at 4 Hz was used as a sensing signal. Among the samples tested, Pt-WO₃/TiO₂ showed the highest sensor response magnitude to NO. The sensor was found to respond consistently and rapidly to change in concentration of NO and NO₂ in the oxygen rich and moist gas mixture at 500 °C. The 90% response and 90% recovery times were as short as less than 5–10 s. The impedance at 4 Hz of the present device was found to vary almost linearly with the logarithm of NO_x (NO or NO₂) concentration from 10 to 570 ppm. Pt-WO₃/TiO₂ showed responses to NO and NO₂ of the same algebraic sign and nearly the same magnitude, while Pt/WO₃ and WO₃/TiO₂ showed higher response to NO than NO₂. The impedance at 4 Hz in the presence of NO for Pt-WO₃/TiO₂ was almost equal at any O₂ concentration examined (1–99%), while in the case of Pt/WO₃ and WO₃/TiO₂ the impedance increased with the oxygen concentration. The features of Pt-WO₃/TiO₂ are favorable as a NO_x sensor that can monitor and control the NO_x concentration in automotive exhaust. The effect of WO₃ loading of Pt-WO₃/ZrO₂-based sensor is studied to discuss the role of surface W-OH sites on the NO_x sensing.     

A sum-of-product neural network (SOPNN)

This paper presents a sum-of-product neural network (SOPNN) structure. The SOPNN can learn to implement static mapping that multilayer neural networks and radial basis function networks normally perform. The output of the neural network has the sum-of-product form ∑Npi=1∏Nvj=1 fij (xj), where xj's are inputs, Nv is the number of inputs, fij( ) is a function generated through network training, and Np is the number of product terms. The function fij(xj) can be expressed as ∑kwijkBjk(xj), where Bjk( ) is a single-variable basis function and Wijk's are weight values. Linear memory arrays can be used to store the weights. If Bjk( ) is a Gaussian function, the new neural network degenerates to a Gaussian function network. This paper focuses on the use of overlapped rectangular pulses as the basis functions. With such basis functions, WijkBjk(xj) will equal either zero or Wijk, and the computation of fij(xj) becomes a simple addition of some retrieved Wijk's. The structure can be viewed as a basis function network with a flexible form for the basis functions. Learning can start with a small set of submodules and have new submodules added when it becomes necessary. The new neural network structure demonstrates excellent learning convergence characteristics and requires small memory space. It has merits over multilayer neural networks, radial basis function networks and CMAC in function approximation and mapping in high-dimensional input space. The technique has been tested for function approximation, prediction of a time series, learning control, and classification.     

Applying the concept of Information Mass in cognizing difficulty of Chinese word strings

The index of cognitive information difficulty (D_info) are measured at the psychological level for the application of learning and design when people are reading and recognizing the Chinese sentences; thus, how to effectively monitor the information difficulty for the purpose of educational learning and design becomes an essential issue. The psychological cognition of the difficulty of sentences experiment is planning to test and verify which one is better of Information Mass (M_info) and Information Quantity. There has been proof that it is suitable for using the concept of M_info in meaningful sentences instead of Information Quantity. We can also measure psychological cognition of the difficulty of sentences from subjects. From our study, we have found that the concept of M_info not only can be use to measure the D_info of Chinese characters but also can be use to measure the D_info of meaningful sentences. There is significant linear relation between D_info of sentences and logarithm of M_info. In other words, we can infer that the D_info of subjects by applying the concept of M_info. The concept of M_info not only can be use to measure the D_info of Chinese characters but also can be use to measure the D_info of meaningful sentences. The authors also establish the qualitative regression model “D_info = a + b*log₂(M_info)”. The concept to apply M_info gives reasonable explanation to the difficulty of Chinese characters and sentences and the model of linear regression serves the function for the reference of educational learning and design.     

稀疏—平坦矩阵信号的重构条件

建立基于凸优化方法重构矩阵信号X=[x₁,…,x_n]的一组充分条件,X具有列稀疏性和平坦性的结构特征,即每个列向量x_j至多具有s个非零分量、同时所有列向量的l₁范数具有相同数值。所采用的矩阵范数是■。工作分两部分,第一部分分别对无观测误差和有观测误差的情况,针对求解■型凸优化问题重构以上类型矩阵信号的方法,建立保障稳定性和鲁棒性的充分条件;第二部分工作建立随机观测情况下观测空间维数的下界,用以保证信号以高概率被正确重构。所得结果优于将重构向量信号的min-l₁方法直接推广到针对矩阵信号的min-l₁方法所得到的结果,并给出数值仿真验证。所针对的信号模型出现在具有定常或缓变包络波形的多输入/多输出雷达及合成孔径雷达等新应用领域,本文工作针对这类应用提供一组实用的信号重构条件。     

多源日光诱导叶绿素荧光产品在中国地区的一致性研究

日光诱导叶绿素荧光(Sun-Induced Chlorophyll Fluorescence,SIF)是表征植被光合的有效探针。基于不同卫星遥感衍生的各种SIF产品已被广泛应用,然而,不同产品在中国区域表现及一致性仍不明晰。以中国生态系统光谱观测网络(China Spec)农田站SIF观测数据为参考评估CSIF、GOSIF、SIF_oco2-005、SIF_LUE(SIF_LUE＿JJ/SIF_LUE＿PK)、SIF₀₀₅等产品,并从不同时空尺度探讨它们的一致性。结果表明：(1)不同产品在4个农田站表现不尽相同,3个农田站存在高估现象。SIF_LUE整体表现较好,GOSIF次之,而后是SIF₀₀₅和SIF_oco2＿005。(2)各产品空间格局具有较高一致性,但其幅度存在差异：年均值方面,SIF_LUE＿PK最大(0.21 W m^-2μm^-1sr^-1<...     

A learning scheme for a fuzzy k-NN rule

The performance of a fuzzy k-NN rule depends on the number k and a fuzzy membership-array W[l,m_R], where l and m_R denote the number of classes and the number of elements in the reference set X_R respectively. The proposed learning procedure consists in iterative finding such k and W which minimize the error rate estimate by the leaving ‘leaving one out‘ method.     

A micromechanical flow sensor for microfluidic applications

We fabricated a microfluidic flow meter and measured its response to fluid flow in a microfluidic channel. The flow meter consisted of a micromechanical plate, coupled to a laser deflection system to measure the deflection of the plate during fluid flow. The 100 /spl mu/m square plate was clamped on three sides and elevated 3 /spl mu/m above the bottom surface of the channel. The response of the flow meter was measured for flow rates, ranging from 2.1 to 41.7 /spl mu/L/min. Several fluids, with dynamic viscosities ranging from 0.8 to 4.5/spl times/10/sup -3/ N/m, were flowed through the channels. Flow was established in the microfluidic channel by means of a syringe pump, and the angular deflection of the plate monitored. The response of the plate to flow of a fluid with a viscosity of 4.5/spl times/10/sup -3/ N/m was linear for all flow rates, while the plate responded linearly to flow rates less than 4.2 /spl mu/L/min of solutions with lower dynamic viscosities. The sensitivity of the deflection of the plate to fluid flow was 12.5/spl plusmn/0.2 /spl mu/rad/(/spl mu/L/min), for a fluid with a viscosity of 4.5/spl times/10/sup -3/ N/m. The encapsulated plate provided local flow information along the length of a microfluidic channel.     

A robustness measure for hypercube networks

Hypercube networks offer a feasible cost-effective solution to parallel computing. Here, a large number of low-cost processors with their own local memories are connected to form an n-cube (B_n) or one of its variants; and the inter-processor communication takes place by message passing instead of shared variables. This paper addresses a constrained two-terminal reliability measure referred to as distance reliability (DR) as it considers the probability that a message can be delivered in optimal time from a given node s to a node t. The problem is equivalent to that of having an operational optimal path (not just any path) between the two nodes. In B_n, the Hamming distance between labels of s and t or H(s, t) determines the length of the optimal path between the two nodes. The shortest distance restriction guarantees optimal communication delay between processors and high link/node utilization across the network. Moreover, it provides a measure for the robustness of symmetric networks. In particular, when H(s, t) = n in B_n, DR will yield the probability of degradation in the diameter, a concept which directly relates to fault-diameter. The paper proposes two schemes to evaluate DR in B_n. The first scheme uses a combinatorial approach by limiting the number of faulty components to (2H(s, t) − 2), while the second outlines paths of length H(s, t) and, then generates a recursive closed-form solution to compute DR. The theoretical results have been verified by simulation. The discrepancy between the theoretical and simulation results is in most cases below 1% and in the worst case 4.6%.     

Potentiometric CO2 gas sensor with lithium phosphorous oxynitride electrolyte

Potentiometric cell, Au/LiCoO₂ 5 m/o Co₃O₄/Li_2.88PO_3.73N_0.14/Li₂CO₃/Au, has been fabricated and investigated for monitoring CO₂ gas. A LiCoO₂–Co₃O₄ mixture was used as the solid-state reference electrode instead of a reference gas. The idea is to keep the lithium activity constant on the reference side using thermodynamic equilibrium at a given temperature. The thermodynamic stability of the reference electrode was studied from the phase stability diagram of Li–Co–C–O system. The Gibb’s free energy of formation of LiCoO₂ was estimated at 500°C from the measured value of the cell emf. The sensors showed good reversibility and fast response toward changing CO₂ concentrations from 200 to 3000 ppm. The emf values were found to follow a logarithmic Nernstian behavior in the 400–500°C temperature range. CH₄ gas did not show any interference effect. Humidity and CO gas decreased the emf values of the sensor slightly. NO and NO₂ gases affect this sensor significantly at low temperatures. However, increased operating temperature seems to reduce the interference.