Dynamic changes in open distributed environments damage the continuous execution of distributed objects. A distributed environment consisting of reconfigurable distributed objects would have higher reliability and ava...
详细信息
Dynamic changes in open distributed environments damage the continuous execution of distributed objects. A distributed environment consisting of reconfigurable distributed objects would have higher reliability and availability, and a greater resilience to unexpected requirement changes. In this paper, we propose a reconfigurable, object model that dynamically changes the object's behavior to fit the current environment by modifying its internal structure. The proposed object consists of communicating concurrent meta-objects. Each meta-object contains functions for adaptation, remote communication, and administration. Therefore, this model provides the required separation of concerns (SoC). Generative communication, which allows meta-object communication, gives the proposed reconfigurable objects the following characteristics: flexibility; ability of allowing a variety of configurations; safety that ensures consistency; and a unification of state preservation and communication. The proposed object model was successfully implemented in a middleware framework called Juice 2.
Distributed systems are some of the most successful structures ever designed for computer users with their undisputed benefits. However, this structure has also introduced several side-effects, most notably unanticipa...
详细信息
Distributed systems are some of the most successful structures ever designed for computer users with their undisputed benefits. However, this structure has also introduced several side-effects, most notably unanticipated runtime events and reconfiguration burdens imposed by environmental changes. In this paper, we discuss a model that enables an object to adapt itself in order to provide the required service by adaptively shifting into the optimal replication strategy. The strategy selection can be done according to environmental conditions in order to address unanticipated events and unpredictable hazards in distributed systems. Therefore, designing an adaptable replication scheme could have a major impact with the growing requirements to support distributed computing to overcome their rapidly growing complexity and to enable their further developments
High Dimensional Model Representation (HDMR) presents the possibility to measure how constant, how constant, how univariate, how bivariate (and so on) a given uniform discrete multivariate data or the multivariate fun...
详细信息
High Dimensional Model Representation (HDMR) presents the possibility to measure how constant, how constant, how univariate, how bivariate (and so on) a given uniform discrete multivariate data or the multivariate function under consideration of HDMR is. Amongst these measurements the most important one is the one for univariance since univariance means additive separability which is the best thing to facilitate the computations in computers. In this paper we define the HDMR's univariance level for a given discrete data. We discuss the possibility of increasing univariance by using certain mappings although the details will not be explicitly given. The basic tool to this end is the recently developed Logarithmic High Dimensional Model Representation and related issues.
This paper is aimed at the illustration of how a recently proposed logarithm based high dimensional model representation (HDMR) works on the functions which have different structures. The main focus of the paper is on...
详细信息
This paper is aimed at the illustration of how a recently proposed logarithm based high dimensional model representation (HDMR) works on the functions which have different structures. The main focus of the paper is on the dominantly multiplicative functions. Since the logarithm converts multiplicativity to additivity what we expect from this new representation is the sufficiency of less number of these new HDMR components than the number of components of plain High Dimensional Model Representation which works well for dominantly additive functions. In implementations we use MuPAD Computer Algebra System to have any desired precision in calculations and to use its symbolic programming nature. We do not use continuous functions instead we take discrete data about the function under consideration. However these data is produced from continuous functions for illustrative purposes. The comparison of the given function values and the values evaluated by truncated Logarithm Based HDMR.
Using traditional methods, it is very difficult to develop high quality, portable software for parallel computers. In particular, parallel software cannot be developed on low cost, sequential computers and then moved ...
详细信息
Using traditional methods, it is very difficult to develop high quality, portable software for parallel computers. In particular, parallel software cannot be developed on low cost, sequential computers and then moved to high performance parallel computers without extensive rewriting and debugging. In the paper, the SFP system being under development at the Institute of informatics Systems is considered. The SFP is aimed to be an interactive visual environment for supporting of functional programming and supercomputing. The input language of the SFP system is a functional language SISAL 3.0 that exposes implicit parallelism through data dependence and guarantees determinate result. The SFP system uses intermediate languages of hierarchical graphs and provides means to write and debug SISAL-programs regardless target architectures as well as to translate the SISAL-programs into optimized imperative programs, appropriate to the target execution platforms.
Unlike wired networks, wireless networks have very dynamic channel properties and the data communication suffers from interference, multi-path fading and noise. Since wireless devices are mobile, factors such as energ...
详细信息
Proteomic study of the synapse has generated an extensive list of molecular components, revealing one of the most complex functional systems currently known to cell biology. While fundamental to neural information pro...
详细信息
This article examines some top concerns in the media industry brought up by media convergence including the need to update news staff, production quality, compensation for multi-platform productions and the legitimacy...
详细信息
Human immunodeficiency virus (HIV) affects millions of people across the globe. Despite the introduction of powerful anti-viral therapies, one factor confounding viral elimination is the ability of HIV to remain laten...
详细信息
ISBN:
(纸本)9812564632
Human immunodeficiency virus (HIV) affects millions of people across the globe. Despite the introduction of powerful anti-viral therapies, one factor confounding viral elimination is the ability of HIV to remain latent within the host genome. Here, we perform a network analysis of the viral reactivation process using human gene expression profiles and curated databases of both human-human and human-HIV protein interactions. Based on this analysis, we report the identification of active pathways in both the latent and early phases of reactivation. These active pathways suggest host functions that are altered and important for HIV pathogenesis.
暂无评论