Joris Borgdorff · UvA-DARE (University of Amsterdam) · 2014
Multiscale models combine knowledge, data, and hypotheses from different scales. Simulating a multiscale model often requires extensive computation. This thesis evaluates distributing these computations, an approach termed distributed multiscale computing (DMC). First, the process of multiscale modelling is examined, in order to describe it in a general and effective way. Then, multiscale models are described with a scale-aware component-based approach, treating them as a set of coupled single scale models. The computational architecture of multiscale applications is then specified with the multiscale modelling language. Such a specification can be analysed for its structural and computational characteristics, using a task graph, and it can be used as the basis for an implementation with Multiscale Coupling Library and Environment 2 (MUSCLE 2). MUSCLE 2 executes multiscale applications on local and distributed machines with a low overhead. As a use case, a model of in-stent restenosis (ISR3D) is described as a set of coupled single scale models, specified with the multiscale modelling language, and implemented and executed with MUSCLE 2. When doing distributed computing, this lead to a decrease in resource consumption under specific circumstances. Five other applications from several domains evaluated DMC, and derived different benefits from it: an increase of simulation speed or a decrease in resource consumption by using heterogeneous machines; or an increase in simulation speed by using more resources altogether. Given these results, DMC is deemed viable for heterogeneous multiscale models and for users with limited local computing resources.
Cited by 0Open access
Özalp Babaoğlu, Alina Sîrbu · CINECA IRIS Institutial research information system (University of Pisa) · 2018
Cognification - the act of transforming ordinary objects or processes into their intelligent counterparts through Data Science and Artificial Intelligence - is a disruptive technology that has been revolutionizing disparate fields ranging from corporate law to medical diagnosis. Easy access to massive data sets, data analytics tools and High-Performance Computing (HPC) have been fueling this revolution. In many ways, cognification is similar to the electrification revolution that took place more than a century ago when electricity became a ubiquitous commodity that could be accessed with ease from anywhere in order to transform mechanical processes into their electrical counterparts. In this paper, we consider two particular forms of distributed computing - Data Centers and HPC systems - and argue that they are ripe for cognification of their entire ecosystem, ranging from top-level applications down to low-level resource and power management services. We present our vision for what "Cognified Distributed Computing" might look like and outline some of the challenges that need to be addressed and new technologies that need to be developed in order to make it a reality. In particular, we examine the role cognification can play in tackling power consumption, resiliency and management problems in these systems. We describe intelligent software-based solutions to these problems powered by on-line predictive models built from streamed real-time data. While we cast the problem and our solutions in the context of large Data Centers and HPC systems, we believe our approach to be applicable to distributed computing in general. We believe that the traditional systems research agenda has much to gain by crossing discipline boundaries to include ideas and techniques from Data Science, Machine Learning and Artificial Intelligence.
Cited by 1Open access
Harry Buhrman, Hein Röhrig · Lecture notes in computer science · 2003
Cited by 93
Jim Farley, Mike Loukides · 1998
From the Publisher:
Distributed computing and Java go together naturally. As the first language designed from the bottom up with networking in mind, Java makes it very easy for computers to cooperate. Even the simplest applet running in a browser is a distributed application, if you think about it. The client running the browser downloads and executes code that is delivered by some other system. But even this simple applet wouldn't be possible without Java's guarantees of portability and security: the applet can run on any platform, and can't sabotage its host.
Of course, when we think of distributed computing, we usually think of applications more complex than a client and server communicating with the same protocol. We usually think in terms of programs that make remote procedure calls, access remote databases, and collaborate with others to produce a single result. Java Distributed Computing discusses how to design and write such applications. It covers Java's RMI (Remote Method Invocation) facility and CORBA, but it doesn't stop there; it tells you how to design your own protocols to build message passing systems and discusses how to use Java's security facilities, how to write multithreaded servers, and more. It pays special attention to distributed data systems, collaboration, and applications that have high bandwidth requirements.
In the future, distributed computing can only become more important. Java Distributed Computing provides a broad introduction to the problems you'll face and the solutions you'll find as you write distributed computing applications.
Topics covered in Java Distributed Computing:
Introduction to Distributed Computing Networking Basics Distributed Objects (Overview of CORBA and RMI) Threads Security Message Passing Systems Distributed Data Systems (Databases) Bandwidth Limited Applications Collaborative Systems
Cited by 104
Jorge Lobo, Jiefei Ma, Alessandra Russo et al. · Lecture notes in computer science · 2012
Cited by 7
Ziad Riad Rahhal, Supervised Dr, Robert Belleman et al. · 2007
Cited by 0
Lei Pan, Lubomir F. Bic, Michael B. Dillencourt et al. · 2004
Sequential computations can benefit from distributed computing through the use of self-migrating threads. On large problems, significant performance improvement is achieved by eliminating disk paging completely, trading off against a modest amount of data communication. The key idea is to have computation and small-sized data move to meet with large-sized data rather than the other way around. Using self-migrating threads with strong mobility provides a flexible way of grouping computations into programs. One of the benefits of this flexibility is ease of applicationlevel programming. Sequential programs are easily augmented into distributed sequential programs for solving large problems, and the distributed sequential programs serve as a good starting point for incremental parallelization. Using our self-migrating threads to implement distributed sequential programs is almost as efficient as using message passing, because we have made thread migration almost as efficient as message passing.
Cited by 0
Distributed Computing Overview · 2007
Introduction The rise of networked workstations and fall of the centralized mainframe has been the most dramatic change in the last two decades of information technology. This shift has put more processing power in the hands of the end-user and distributed hardware resources throughout the enterprise. No longer the domain of raised floors and data centers, processing power now resides on desktops, workgroup servers, and minis. This shift first involved hardware; the current challenge is to develop the software infrastructure to make use of these now distributed resources. As networks of computing resources have become prevalent, the concept of distributing related processing among multiple resources has become increasingly viable and desirable. Over the years, several methods have evolved to enable this distribution, ranging from simplistic data sharing to advanced systems supporting a multitude of services. This paper presents an overview of the means used to enable distribut
Cited by 0
New Electronics · 2019
Data is the key to any business and Intertrust's suite of computing products looks to better manage and understand it, as CEO Talal Shamoon explains to New Electronics
Cited by 0
Bode, Arndt 1948-, EDMCC 1991 München · Lecture notes in computer science · 1991
Cited by 18
Lorenzo Natale · 2020
Cited by 0
Nancy Lynch · 2007
No abstract available.
Cited by 5
Günther Sawitzki · heiDOK (Heidelberg University) · 2000
Demonstration of NetWork, a system for asynchronous distributed computing in a non-guarantied environment.
Cited by 0Open access
Sven Ubik · 2008
Cited by 0
J. B. Sidney, J. Urrutia · 2008
Since the pioneering research of Cook [1] and Karp [3] in computational complexity, an enormous amount of research has been directed toward establishing
Cited by 0
Prof Dr, Roger Wattenhofer, Advisor Yvonne et al. · 2007
Summer 2007First of all I especially would like to thank my parents for their great support during my studies. Then I have to thank my brothers and friends which supported me greatly the last few months. Furthermore I must thank my advisor and my supervising professor who were responsible for a very beneficial research atmosphere and provided a great support. Word of Mouth
Cited by 0
Ali Khalesi, Petros Elia · 2024
The work considers the N-server distributed computing scenario with K users requesting functions that are linearly-decomposable over an arbitrary basis of L real (potentially non-linear) subfunctions. In our problem, the aim is for each user to receive their function outputs, allowing for reduced reconstruction error (distortion) ϵ, reduced computing cost (γ; the fraction of subfunctions each server must compute), and reduced communication cost (δ; the fraction of users each server must connect to). For any given set of K requested functions-which is here represented by a coefficient matrix F ∈ R K×L-our problem is made equivalent to the open problem of sparse matrix factorization that seeks-for a given parameter T , representing the number of shots for each server-to minimize the reconstruction distortion 1 KL ∥F − DE∥ 2 F over all δ-sparse and γ-sparse matrices D ∈ R K×N T and E ∈ R N T ×L. With these matrices respectively defining which servers compute each subfunction, and which users connect to each server, we here design our D, E by designing tessellated-based and SVD-based fixed support matrix factorization methods that first split F into properly sized and carefully positioned submatrices, which we then approximate and then decompose into properly designed submatrices of D and E. For the zero-error case and under basic dimensionality assumptions, the work reveals achievable computationvs-communication corner points (γ, δ) which, for various cases, are proven optimal over a large class of D, E by means of a novel tessellations-based converse. Subsequently, for large N , and under basic statistical assumptions on F, the average achievable error ϵ is concisely expressed using the incomplete first moment of the standard Marchenko-Pastur distribution, where this performance is shown to be optimal over a large class of D and E. In the end, the work also reveals that the overall achieved gains over baseline methods are unbounded.
Cited by 1Open access
Jean-luc Geering, Prof Dr, Roger Wattenhofer et al. · 2012
A game is both an interesting and a complex application. Having a shared state makes it a challenging undertaking to develop it in a peer-to-peer setting. XP2Pilot is an implementation of such a game. It is a fully decentralized version of XPilot-NG, built on top of Pulsar. Acknowledgments First of all, I would like to thank Prof. Dr. Wattenhofer for giving me the opportunity to work on this most interesting and challenging project. During this truly unique experience, I benefitted not only on a technical but also on a personal level. Stefan Schmid and Thomas Locher, my faithful advisors, thank you for not breaking my e-bike when you took it for a ride. Special thanks to Remo Meier who allowed me to use Pulsar and was always quick to respond to my numerous questions. 1 Cordial thanks to the open-source communities. They make the world a better place. At least the virtual one. To the Swiss taxpayers, thank you for financing my education. Keep up the good work, the SFITZ 2 is such an awesome institution.
Cited by 0
Matt Farrens · ACM Computing Surveys · 1996
No abstract available.
Cited by 4
Raphael Eidenbenz, Prof Dr, Roger Wattenhofer et al. · 2012
Abstract — This paper attends to the problem of a mechanism designer seeking to influence the outcome of a strategic game based on her creditability. The mechanism designer offers additional payments to the players depending on their mutual choice of strategies in order to steer them to certain decisions. If the players trust her, they will select a newly profitable strategy. Of course, the mechanism designer aims at spending as little as possible and yet implement her desired outcome. We present several algorithms for this optimization problem both for optimal solutions and approximations thereof. We analyze singleton target strategy profiles and target strategy profile sets. Furthermore, the paper shows how a bankrupt mechanism designer can decide efficiently whether strategy profiles can be implemented at no cost at all. Finally risk-averse players and dynamic games are examined. I.
Cited by 0