Meta level component-based framework for distributed computing systems

Adaptability for distributed object-oriented enterprise frameworks in multimedia technology is a critical mission for system evolution. Today, building adaptive services is a complex task due to lack of adequate framework support in the distributed computing systems. In this paper, we propose a Metalevel Component-Based Framework which uses distributed computing design patterns as components to develop an adaptable pattern-oriented framework for distributed computing applications. We describe our approach of combining a meta-architecture with a pattern-oriented framework, resulting in an adaptable framework which provides a mechanism to facilitate system evolution. This approach resolves the problem of dynamic adaptation in the framework, which is encountered in most distributed multimedia applications. The proposed architecture of the pattern-oriented framework has the abilities to dynamically adapt new design patterns to address issues in the domain of distributed computing and they can be woven together to shape the framework in future. © 2011 Springer Science+Business Media B.V.

Message passing for distributed optimisation of power allocation with renewable resources

The increase in renewable energy generators introduced into the electricity grid is putting pressure on its stability and management as predictions of renewable energy sources cannot be accurate or fully controlled. This, with the additional pressure of fluctuations in demand, presents a problem more complex than the current methods of controlling electricity distribution were designed for. A global approximate and distributed optimisation method for power allocation that accommodates uncertainties and volatility is suggested and analysed. It is based on a probabilistic method known as message passing [1], which has deep links to statistical physics methodology. This principled method of optimisation is based on local calculations and inherently accommodates uncertainties; it is of modest computational complexity and provides good approximate solutions.We consider uncertainty and fluctuations drawn from a Gaussian distribution and incorporate them into the message-passing algorithm. We see the effect that increasing uncertainty has on the transmission cost and how the placement of volatile nodes within a grid, such as renewable generators or consumers, effects it.

Effect of Rayleigh-scattering distributed feedback in multi-wavelength and tunable Raman fibre lasers

We have demonstrated that a random distributed feedback based on the Rayleigh scattering provides very flat power-versus-wavelength characteristics both in tunable and multiwavelength ultra-long fibre lasers. © 2011 Optical Society of America.

Intensity dynamics and statistical properties of random distributed feedback fiber laser

We present first experimental investigation of fast-intensity dynamics of random distributed feedback (DFB) fiber lasers. We found that the laser dynamics are stochastic on a short time scale and exhibit pronounced fluctuations including generation of extreme events. We also experimentally characterize statistical properties of radiation of random DFB fiber lasers. We found that statistical properties deviate from Gaussian and depend on the pump power.

Random distributed feedback fiber laser

I will overview our recent results on ultra-long lasers and will discuss the concept of a fiber laser with an open cavity that operates using random distributed feedback provided by Rayleigh scattering amplified through the Raman effect. © 2011 Optical Society of America.

Ultrahigh step-up dc-dc converter for distributed generation by three Degrees of Freedom (3DoF) approach

This paper proposes a novel dc-dc converter topology to achieve an ultrahigh step-up ratio while maintaining a high conversion efficiency. It adopts a three degree of freedom approach in the circuit design. It also demonstrates the flexibility of the proposed converter to combine with the features of modularity, electrical isolation, soft-switching, low voltage stress on switching devices, and is thus considered to be an improved topology over traditional dc-dc converters. New control strategies including the two-section output voltage control and cell idle control are also developed to improve the converter performance. With the cell idle control, the secondary winding inductance of the idle module is bypassed to decrease its power loss. A 400-W dc-dc converter is prototyped and tested to verify the proposed techniques, in addition to a simulation study. The step-up conversion ratio can reach 1:14 with a peak efficiency of 94% and the proposed techniques can be applied to a wide range of high voltage and high power distributed generation and dc power transmission.

Benchmarking of distributed computing engines spark and GraphLab for big data analytics

In this paper we evaluate and compare two representativeand popular distributed processing engines for large scalebig data analytics, Spark and graph based engine GraphLab. Wedesign a benchmark suite including representative algorithmsand datasets to compare the performances of the computingengines, from performance aspects of running time, memory andCPU usage, network and I/O overhead. The benchmark suite istested on both local computer cluster and virtual machines oncloud. By varying the number of computers and memory weexamine the scalability of the computing engines with increasingcomputing resources (such as CPU and memory). We also runcross-evaluation of generic and graph based analytic algorithmsover graph processing and generic platforms to identify thepotential performance degradation if only one processing engineis available. It is observed that both computing engines showgood scalability with increase of computing resources. WhileGraphLab largely outperforms Spark for graph algorithms, ithas close running time performance as Spark for non-graphalgorithms. Additionally the running time with Spark for graphalgorithms over cloud virtual machines is observed to increaseby almost 100% compared to over local computer clusters.

Reach extension of high-capacity LR-PONs in chain configuration of distributed amplifier structure

We show 100-Gb/s single-channel transmission in LR-PONs with at least 512 way split and up to 160 km total distance is feasible by means of "chained" cable with amplifier solutions, and appropriate FIR filter designs. © 2015 OSA.

Random distributed feedback fiber laser

Researchers conducted investigations to demonstrate the advantages of random distributed feedback fiber laser. Random lasers had advantages, such as simple technology that did not require a precise microcavity and low production cost. The properties of their output radiation were special in comparison to those of conventional lasers and they were characterized by complex features in the spatial, spectral, and time domains. The researchers demonstrated a new type of one-dimensional laser with random distributed feedback based on Rayleigh scattering (RS) that was presented in any transparent glass medium due to natural inhomogeneities of refractive index. The cylindrical fiber waveguide geometry provided transverse confinement, while the cavity was open in the longitudinal direction and did not include any regular point-action reflectors.

RIN transfer in 2nd-order distributed amplification with ultralong fiber lasers

We investigate numerically the effect of ultralong Raman laser fiber amplifier design parameters, such as span length, pumping distribution and grating reflectivity, on the RIN transfer from the pump to the transmitted signal. Comparison is provided to the performance of traditional second-order Raman amplified schemes, showing a relative performance penalty for ultralong laser systems that gets smaller as span length increases. We show that careful choice of system parameters can be used to partially offset such penalty. © 2010 Optical Society of America.

Raman fiber lasers with a random distributed feedback based on Rayleigh scattering

We demonstrate lasing based on a random distributed feedback due to the Raman amplified Rayleigh backscattering in different types of cavities with and without conventional point-action reflectors. Quasistationary generation of a narrowband spectrum is achieved despite the random nature of the feedback. The generated spectrum is localized at the reflection or gain spectral maxima in schemes with and without point reflectors, respectively. The length limit for a conventional cavity and the minimal pump power required for the lasing based purely on a random distributed feedback are determined. © 2010 The American Physical Society.

Eigen-Adaptation and Distributed Representation of 2-D Phase, Energy, Scale and Orientation in Spatial Vision

Distributed representations (DR) of cortical channels are pervasive in models of spatio-temporal vision. A central idea that underpins current innovations of DR stems from the extension of 1-D phase into 2-D images. Neurophysiological evidence, however, provides tenuous support for a quadrature representation in the visual cortex, since even phase visual units are associated with broader orientation tuning than odd phase visual units (J.Neurophys.,88,455–463, 2002). We demonstrate that the application of the steering theorems to a 2-D definition of phase afforded by the Riesz Transform (IEEE Trans. Sig. Proc., 49, 3136–3144), to include a Scale Transform, allows one to smoothly interpolate across 2-D phase and pass from circularly symmetric to orientation tuned visual units, and from more narrowly tuned odd symmetric units to even ones. Steering across 2-D phase and scale can be orthogonalized via a linearizing transformation. Using the tiltafter effect as an example, we argue that effects of visual adaptation can be better explained by via an orthogonal rather than channel specific representation of visual units. This is because of the ability to explicitly account for isotropic and cross-orientation adaptation effect from the orthogonal representation from which both direct and indirect tilt after-effects can be explained.

Learning user and product distributed representations using a sequence model for sentiment analysis

In product reviews, it is observed that the distribution of polarity ratings over reviews written by different users or evaluated based on different products are often skewed in the real world. As such, incorporating user and product information would be helpful for the task of sentiment classification of reviews. However, existing approaches ignored the temporal nature of reviews posted by the same user or evaluated on the same product. We argue that the temporal relations of reviews might be potentially useful for learning user and product embedding and thus propose employing a sequence model to embed these temporal relations into user and product representations so as to improve the performance of document-level sentiment analysis. Specifically, we first learn a distributed representation of each review by a one-dimensional convolutional neural network. Then, taking these representations as pretrained vectors, we use a recurrent neural network with gated recurrent units to learn distributed representations of users and products. Finally, we feed the user, product and review representations into a machine learning classifier for sentiment classification. Our approach has been evaluated on three large-scale review datasets from the IMDB and Yelp. Experimental results show that: (1) sequence modeling for the purposes of distributed user and product representation learning can improve the performance of document-level sentiment classification; (2) the proposed approach achieves state-of-The-Art results on these benchmark datasets.

Internet computing with distributed software agent

The phenomenonal growth of the Internet has connected us to a vast amount of computation and information resources around the world. However, making use of these resources is difficult due to the unparalleled massiveness, high communication latency, share-nothing architecture and unreliable connection of the Internet. In this dissertation, we present a distributed software agent approach, which brings a new distributed problem-solving paradigm to the Internet computing researches with enhanced client-server scheme, inherent scalability and heterogeneity. Our study discusses the role of a distributed software agent in Internet computing and classifies it into three major categories by the objects it interacts with: computation agent, information agent and interface agent. The discussion of the problem domain and the deployment of the computation agent and the information agent are presented with the analysis, design and implementation of the experimental systems in high performance Internet computing and in scalable Web searching. ^ In the computation agent study, high performance Internet computing can be achieved with our proposed Java massive computation agent (JAM) model. We analyzed the JAM computing scheme and built a brutal force cipher text decryption prototype. In the information agent study, we discuss the scalability problem of the existing Web search engines and designed the approach of Web searching with distributed collaborative index agent. This approach can be used for constructing a more accurate, reusable and scalable solution to deal with the growth of the Web and of the information on the Web. ^ Our research reveals that with the deployment of the distributed software agent in Internet computing, we can have a more cost effective approach to make better use of the gigantic scale network of computation and information resources on the Internet. The case studies in our research show that we are now able to solve many practically hard or previously unsolvable problems caused by the inherent difficulties of Internet computing. ^

Engineering access control for distributed enterprise applications

Access control (AC) is a necessary defense against a large variety of security attacks on the resources of distributed enterprise applications. However, to be effective, AC in some application domains has to be fine-grain, support the use of application-specific factors in authorization decisions, as well as consistently and reliably enforce organization-wide authorization policies across enterprise applications. Because the existing middleware technologies do not provide a complete solution, application developers resort to embedding AC functionality in application systems. This coupling of AC functionality with application logic causes significant problems including tremendously difficult, costly and error prone development, integration, and overall ownership of application software. The way AC for application systems is engineered needs to be changed. ^ In this dissertation, we propose an architectural approach for engineering AC mechanisms to address the above problems. First, we develop a framework for implementing the role-based access control (RBAC) model using AC mechanisms provided by CORBA Security. For those application domains where the granularity of CORBA controls and the expressiveness of RBAC model suffice, our framework addresses the stated problem. ^ In the second and main part of our approach, we propose an architecture for an authorization service, RAD, to address the problem of controlling access to distributed application resources, when the granularity and support for complex policies by middleware AC mechanisms are inadequate. Applying this architecture, we developed a CORBA-based application authorization service (CAAS). Using CAAS, we studied the main properties of the architecture and showed how they can be substantiated by employing CORBA and Java technologies. Our approach enables a wide-ranging solution for controlling the resources of distributed enterprise applications. ^