FREP: Energy proportionality for disk storage using replication

Energy saving has become a crucial concern in datacenters as several reports predict that the anticipated energy costs over a three year period will exceed hardware acquisition. In particular, saving energy for storage is of major importance as storage devices (and cooling them off) may contribute over 25 percent of the total energy consumed in a datacenter. Recent work introduced the concept of energy proportionality and argued that it is a more relevant metric than just energy saving as it takes into account the tradeoff between energy consumption and performance. In this paper, we present a novel approach, called FREP (Fractional Replication for Energy Proportionality), for energy management in large datacenters. FREP includes a replication strategy and basic functions to enable flexible energy management. Specifically, our method provides performance guarantees by adaptively controlling the power states of a group of disks based on observed and predicted workloads. Our experiments using a set of real and synthetic traces show that FREP dramatically reduces energy requirements with a minimal response time penalty.     

Extractors for varieties

We study the task of randomness extraction from sources that are distributed uniformly on an unknown algebraic variety. In other words, we are interested in constructing a function (an extractor) whose output is close to uniform even if the input is drawn uniformly from the set of solutions of an unknown system of low degree polynomials. This problem generalizes the problem of extraction from affine sources which has drawn a considerable amount of interest lately. We present two constructions of explicit extractors for varieties. The first works for varieties of any size (including one-dimensional varieties or curves) and requires field size that is exponential in the overall dimension of the space. Our second extractor allows the field size to be polynomial in the degree of the equations defining the variety, but works only for varieties whose size is at least the square root of the total size of the space.     

Naturally occurring mastitis disrupts developmental competence of bovine oocytes

We examined the effects of naturally occurring mastitis on bovine oocyte developmental competence in vitro. Specifically, we investigated the effects of intramammary infection on the ovarian pool of oocytes (i.e., follicle-enclosed oocytes) and their ability to undergo in vitro maturation, fertilization, and further development to the blastocyst stage. Culled Holstein cows (n = 50) from 9 commercial dairy farms in Israel were allotted to 3 groups according to somatic cell count (SCC) records of the last 3 monthly milk tests as well as of quarter samples collected before slaughter: (1) low SCC (n = 7), (2) medium SCC (n = 16), or (3) high SCC (n = 27). Means of SCC values differed among low-, medium-, and high-SCC groups: 148,000, 311,000 and 1,813,000 cell/mL milk, respectively. Milk yield and days in milk did not differ among the 3 groups. Bacterial isolates included coagulase-negative staphylococci, Escherichia coli, Streptococcus dysgalactiae, or no bacteria found. Ovaries were collected at the abattoir and brought to the laboratory. Cumulus oocyte complexes were recovered separately from each cow and subjected individually to in vitro maturation and fertilization, followed by 8 d in culture. The number of aspirated oocytes did not differ among groups, with a range of 17 to 21 oocytes per cow. The proportion of oocytes that cleaved into 2- to 4-cell-stage embryos (86.1 ± 3.4%) did not differ among groups. In contrast, mean percentages of embryos developed to the blastocyst stage on d 7 and 8 after fertilization were less in both medium- and-high SCC groups than in the low-SCC group (5.6 ± 2.3 and 4.1 ± 1.8 vs. 18.1 ± 4.6%, respectively). Additional analysis indicated that cleavage and blastocyst-formation rates did not differ among the bacterial types in the low-, medium-, and high-SCC groups. These are the first results to demonstrate that naturally occurring mastitis disrupts the developmental competence of the ovarian pool of oocytes, (i.e., oocytes at the germinal vesicle stage). The disruption was associated with elevation of SCC rather than bacterial type. The results may provide a partial explanation for the low fertility of cows that have contracted mastitic pathogens before insemination.     

Virus concentration using hollow fiber membranes—II

An asymmetric polysulphone hollow fiber membrane permeator was used to concentrate polio I virus from 5 and 50-L of deionized- and tap-water. The objective of the study was to evaluate the effect of several independent variables—membrane type and morphology, initial virus concentration, initial feed volume, feed water type and backwash recovery of virus—on process performance. Also, it was the purpose of this study to compare performance for the asymmetric polysulphone and semidense symmetric cellulose acetate hollow fiber membranes (Belfort, Rotem and Katzenelson, 1975). From the results presented here and from a general chemical and biological stability point of view, the polysulphone membranes were superior for virus concentration. The module tested here proved to be versatile and continuously reusable. Finally, in no case and under all conditions studied was polio I virus detected in the permeate.     

Multiscale Quantum-Classical Simulations of Enzymes

Enzymes are remarkably efficient catalysts evolved to perform well-defined and highly specific chemical transformations. Studying the nature of enzymatic rate enhancements is highly important from several aspects, including the rational design of synthetic catalysts and transition-state inhibitors. Herein, we describe recent progress in our group in the development of multiscale simulation methods and their application to several enzyme systems. In particular, we describe the use of combined quantum mechanics/molecular mechanics (QM/MM) methods in classical and quantum simulations. The development of various novel path-integral methods is reviewed. These methods are tailor-made for enzyme systems, where only a few degrees of freedom involved in the chemistry need to be quantized. The application of the hybrid QM/MM quantum-classical simulation approach to three case studies is presented. The first case involves proton transfer in nitroalkane oxidase, where the enzyme employs tunneling as a catalytic fine-tuning tool. The second case presented involves orotidine 5′-monophosphate decarboxylase, where multidimensional free energy simulations together with kinetic isotope effects are combined in the study of the reaction mechanism. Finally, we discuss the monoterpene cyclase bornyl diphosphate synthase, where non-statistical dynamics is a key component in enzyme function.     

需求推动创新DSP与FPGA协力优化摄像监视系统

摄像监视系统现在具有更多的功能,除了提供先进的压缩技术如MPEG-4和H.264外,还配备了如图像稳定、全景摄像和视频运动检测算法等,是传统模拟系统强有力的替代方案.本文将讨论这些新技术的优点以及它们在用数字信号处理器(DSP)和FPGA协处理器平台上的优化实现.     

A bicriteria approach to minimize maximal lateness and resource consumption for scheduling a single machine

We extend the classical single-machine maximal lateness scheduling problem to the case where the job processing times are controllable by allocating a continuous and nonrenewable resource to the processing operations. Our aim is to construct an efficient trade-off curve between maximal lateness and total resource consumption using a bicriteria approach. We present a polynomial time algorithm that constructs this trade-off curve assuming a specified general type of convex decreasing resource consumption function. We illustrate the algorithm with a numerical example.     

A bicriterion approach to time/cost trade-offs in scheduling with convex resource-dependent job processing times and release dates

This paper presents a bicriterion analysis of time/cost trade-offs for the single-machine scheduling problem where both job processing times and release dates are controllable by the allocation of a continuously nonrenewable resource. Using the bicriterion approach, we distinguish between our sequencing criterion, namely the makespan, and the cost criterion, the total resource consumed, in order to construct an efficient time/cost frontier. Although the computational complexity of the problem of constructing this frontier remains an open question, we show that the optimal job sequence is independent of the total resource being used; thereby we were able to reduce the problem to a sequencing one. We suggest an exact dynamic programming algorithm for solving small to medium sizes of the problem, while for large-scale problems we present some heuristic algorithms that turned out to be very efficient. Five different special cases that are solvable by using polynomial time algorithms are also presented.