We present BlowFish, a distributed data store that admits a smooth tradeoff between storage and performance for point queries. What makes BlowFish unique is its ability to navigate along this tradeoff curve efficiently at fine-grained time scales with low computational overhead. Achieving a smooth and dynamic storage-performance tradeoff enables a wide range of applications. We apply BlowFish to several such applications from real-world production clusters: (i) as a data recovery mechanism during failures: in practice, BlowFish requires 5.4× lower bandwidth and 2.5× lower repair time compared to state-of-the-art erasure codes, while reducing the storage cost of replication from 3× to 1.9×; and (ii) data stores with spatially-skewed and time-varying workloads (e.g., due to object popularity and/or transient failures): we show that navigating the storage-performance tradeoff achieves higher system-wide utility (e.g., throughput) than selectively caching hot objects.
A statistical Ritchie-Knott-Rice (RKR) [1 ] model for brittle fracture is considered for an FGM containing a slender notch. The FMG is modeled as linear elastic, with its strength described by two-parameter Weibull statistics. The Young's modulus is assumed to vary either linearly or sigmoidally. A compact tension (C(T)) fracture mechanics specimen is analyzed via the finite element method, considering the effect of modulus variation on the near-tip stress state. Results can be characterized by the stress intensity, K. For spatially constant Weibull parameters, the RKR model is used to predict the expected fracture toughness, K Φ , i.e., the K at which the first flaw failure occurs with probability Φ. For sufficiently high Weibull modulus, the failure occurs essentially at the notch tip. For sufficiently low Weibull modulus (m < 4), K Φ for an FGM is found to vary up to 25% from that of a homogeneous body.
We present CellIQ, a real-time cellular network analytics system that supports rich and sophisticated analysis tasks. CellIQ is motivated by the lack of support for realtime analytics or advanced tasks such as spatio-temporal traffic hotspots and handoff sequences with performance problems in state-of-the-art systems, and the interest in such tasks by network operators. CellIQ represents cellular network data as a stream of domain specific graphs, each from a batch of data. Leveraging domain specific characteristics--the spatial and temporal locality of cellular network data--CellIQ presents a number of optimizations including geo-partitioning of input data, radius-based message broadcast, and incremental graph updates to support efficient analysis. Using data from a live cellular network and representative analytic tasks, we demonstrate that CellIQ enables fast and efficient cellular network analytics--compared to an implementation without cellular specific operators, CellIQ is 2× to 5× faster.
Background Previous models of good continuation [e.g., Williams & Jacobs 95] make the first-order Markov assumption, i.e., if we parametrize a curve by arc length t, the tangent direction of the contour at t+1 only depends on the tangent at t. The goal of this study is to use human-marked boundary contours in a large database of natural images to empirically determine the validity of this model. Methods Experiment 1: We measure the distribution of lengths of contours segmented at local curvature maxima. If the first-order Markov assumption holds, the lengths of the segments would have an exponential distribution. Experiment 2: We evaluate higher-order Markov models, in which the tangent direction of a contour at t+1 depends on the tangent at t and the tangents at the same location t of this contour at coarser scales. This is done by empirically measuring the information gain when the order of our model increases, i.e., the mutual information between the tangent at t+1 and the tangent at t at scale s conditioned on all the tangents at t at scales finer than s. Results Experiment 1: We observe a power law, instead of an exponential law, in the distribution of the contour segment length. The probability is inversely proportional to the square of segment length. The power law justifies the intuition that contours are multi-scale in nature; the first-order Markov assumption is shown to be empirically invalid. Experiment 2: The information gain shows that coarser scales contain a significant amount of information (17% of the base scale). We accordingly propose a multi-scale algorithm for contour completion, which uses higher-order Markov models. Completion is done in a coarse-to-fine manner. Conclusion Any algorithm for contour processing has to be intrinsically multi-scale. Higher-order Markov models exploit information across scales and lead to an efficient algorithm for multi-scale contour completion.
