Filters








400 Hits in 0.94 sec

Pascal Felber : Siliziumritter alter Garde

Pierre-Yves Frei
2004
Pascal Felber verbrachte viel Zeit damit zu verstehen, was «darunter» abläuft.  ...  Eine Lösung könnte für Pascal Felber die Auflösung der Hierarchien bringen. Eine Revolution also.  ... 
doi:10.5169/seals-552754 fatcat:6a52ioslkfguxogkmflxeuzbcq

Elastic Transactions [chapter]

Pascal Felber, Vincent Gramoli, Rachid Guerraoui
2009 Lecture Notes in Computer Science  
This paper presents elastic transactions, an appealing alternative to traditional transactions, in particular to implement search structures in shared memory multicore architectures. Upon conflict detection, an elastic transaction might drop what it did so far within a separate transaction that immediately commits, and resume its computation within a new transaction which might itself be elastic. We present the elastic transaction model and an implementation of it, then we illustrate its
more » ... lustrate its simplicity and performance on various concurrent data structures, namely double-ended queue, hash table, linked list, and skip list. Elastic transactions outperform classical ones on various workloads, with an improvement of 35% on average. They also exhibit competitive performance compared to lock-based techniques and are much simpler to program with than lock-free alternatives. $ A preliminary extended abstract of this work has been published in the proceedings of DISC 2009 [15], the current version extends it by generalizing the model, applying it to additional data structures, and comparing it against existing synchronization alternatives.
doi:10.1007/978-3-642-04355-0_12 fatcat:zwx4uz4g7ngd7icnvguae5by4q

Network-Friendly Gossiping [chapter]

Sabina Serbu, Étienne Rivière, Pascal Felber
2009 Lecture Notes in Computer Science  
The emergence of large-scale distributed applications based on many-to-many communication models, e.g., broadcast and decentralized group communication, has an important impact on the underlying layers, notably the Internet routing infrastructure. To make an eective use of network resources, protocols should both limit the stress (amount of messages) on each infrastructure entity like routers and links, and balance as much as possible the load in the network. Most protocols use
more » ... se application-level metrics such as delays to improve eciency of content dissemination or routing, but the extend to which such applicationcentric optimizations help reduce and balance the load imposed to the infrastructure is unclear. In this paper, we elaborate on the design of such network-friendly protocols and associated metrics. More specically, we investigate random-based gossip dissemination. We propose and evaluate dierent ways of making this representative protocol network-friendly while keeping its desirable properties (robustness and low delays). Simulations of the proposed methods using synthetic and real network topologies convey and compare their abilities to reduce and balance the load while keeping good performance.
doi:10.1007/978-3-642-05118-0_45 fatcat:a2a3bu643jcmnnop6gejta5uvu

A Wait-Free Universal Construct for Large Objects [article]

Andreia Correia, Pedro Ramalhete, Pascal Felber
2019 arXiv   pre-print
Authors' addresses: Andreia Correia, University of Neuchatel, andreia.veiga@unine.ch; Pedro Ramalhete, Cisco Systems, pramalhe@gmail.com; Pascal Felber, University of Neuchatel, pascal.felber@unine.ch.  ... 
arXiv:1911.01676v1 fatcat:y6cfg4do3vhijmwiddteolhy2e

Elastic transactions

Pascal Felber, Vincent Gramoli, Rachid Guerraoui
2017 Journal of Parallel and Distributed Computing  
This paper presents elastic transactions, an appealing alternative to traditional transactions, in particular to implement search structures in shared memory multicore architectures. Upon conflict detection, an elastic transaction might drop what it did so far within a separate transaction that immediately commits, and resume its computation within a new transaction which might itself be elastic. We present the elastic transaction model and an implementation of it, then we illustrate its
more » ... lustrate its simplicity and performance on various concurrent data structures, namely double-ended queue, hash table, linked list, and skip list. Elastic transactions outperform classical ones on various workloads, with an improvement of 35% on average. They also exhibit competitive performance compared to lock-based techniques and are much simpler to program with than lock-free alternatives. $ A preliminary extended abstract of this work has been published in the proceedings of DISC 2009 [15], the current version extends it by generalizing the model, applying it to additional data structures, and comparing it against existing synchronization alternatives.
doi:10.1016/j.jpdc.2016.10.010 fatcat:tanoigl42jhyhdhjutdkcy75fy

Semi-automatic Parallelization of Java Applications [chapter]

Pascal A. Felber
2003 Lecture Notes in Computer Science  
Some types of time-consuming computations are naturally parallelizable. To take advantage of parallel processing, however, applications must be explicitly programmed to use specific libraries that share the workload among multiple (generally distributed) processors. In this paper, we present a set of Java tools that allow us to parallelize some types of computationally-intensive Java applications a posteriori, even when the source code of these applications is not available. Our tools operate
more » ... Our tools operate using techniques based on bytecode transformation, code migration, and distributed parallel method executions.
doi:10.1007/978-3-540-39964-3_86 fatcat:uonksq3ewbdkdmqxyytivaloyy

Advanced concurrency control in Java

Pascal Felber, Michael K. Reiter
2002 Concurrency and Computation  
The benchmarking application and an implementation of JCF are available at http://www.belllabs.com/user/felber/atomic/.  ... 
doi:10.1002/cpe.635 fatcat:w2na2murhvc2rc2md77vylhjvy

System support for object groups

Rachid Guerraoui, Pascal Felber, Benoît Garbinato, Karim Mazouni
1998 SIGPLAN notices  
This paper draws several observations from our experiences in building support for object groups. These observations actually go beyond our experiences and may apply to many other developments of object based distributed systems. Our first experience aimed at building support for Smalltalk object replication using the Isis process group toolkit. It was quite easy to achieve group transparency but we were confronted with a strong mismatch between the rigidity of the process group model and the
more » ... oup model and the flexible nature of object interactions. Consequently, we decided to build our own object oriented protocol framework, specifically dedicated to support object groups (instead of using a process group toolkit). We built our framework in such a way that basic distributed protocols, such as failure detection and multicasts, are considered as first class entities, directly accessible to the programmers. To achieve flexible and dynamic protocol composition, we had to go beyond inheritance and objectify distributed algorithms. Our second experience consisted in building a CORBA service aimed at managing group of objects written on different languages and running on different platforms. This experience revealed a mismatch between the asynchrony of group protocols and the synchrony of standard CORBA interaction mechanisms, which limited the portability of our CORBA object group service. We restricted the impact of this mismatch by encapsulating asynchrony issues inside a specific messaging sub-service. We dissect the cost of object group transparency in our various implementations, and we point out the recurrent sources of overheads, namely message indirection, marshaling/unmarshaling and strong consistency.
doi:10.1145/286942.286961 fatcat:xbeg4cezmjep7izha6zs7qvga4

HyPeer: Structured overlay with flexible-choice routing

Sabina Serbu, Pascal Felber, Peter Kropf
2011 Computer Networks  
The various designs of peer-to-peer distributed hash tables (DHTs) differ mostly in their performance and the properties they provide. They usually rely on a single greedy routing strategy aiming at achieving short communication paths. In this paper, we propose a new peer-to-peer DHT overlay design, HYPEER, that uses the "flexible-choice routing" principle to support multiple routing strategies, each driven by a different objective: fault tolerance, load balance, low latency. Our overlay is
more » ... Our overlay is loosely based on a hypercube structure providing redundant paths that can be selectively followed at runtime according to the desired goal. HYPEER approximates a hypercube by extending a ring-based DHT overlay and controlling the placement of nodes when they join the network. Experimental evaluation confirms that HYPEER succeeds in improving the routing performance according to our three design objectives, while only slightly increasing the average path length.
doi:10.1016/j.comnet.2010.09.006 fatcat:4m7nsylkbzgwrjchyvqsosrp4y

SplayNet: Distributed User-Space Topology Emulation [chapter]

Valerio Schiavoni, Etienne Rivière, Pascal Felber
2013 Lecture Notes in Computer Science  
Network emulation allows researchers to test distributed applications on diverse topologies with fine control over key properties such as delays, bandwidth, congestion, or packet loss. Current approaches to network emulation require using dedicated machines and low-level operating system support. They are generally limited to one user deploying a single topology on a given set of nodes, and they require complex management. These constraints restrict the scope and impair the uptake of network
more » ... ptake of network emulation by designers of distributed applications. We propose a set of novel techniques for network emulation that operate only in user-space without specific operating system support. Multiple users can simultaneously deploy several topologies on shared physical nodes with minimal setup complexity. A modular network model allows emulating complex topologies, including congestion at inner routers and links, without any centralized orchestration nor dedicated machine. We implement our user-space network emulation mechanisms in SplayNet, as an extension of an open-source distributed testbed. Our evaluation with a representative set of applications and topologies shows that SplayNet provides accuracy comparable to that of low-level systems based on dedicated machines, while offering better scalability and ease of use.
doi:10.1007/978-3-642-45065-5_4 fatcat:hnny6lznn5cxrhyhvcx5idcqr4

Cooperative Content Distribution: Scalability Through Self-Organization [chapter]

Pascal Felber, Ernst W. Biersack
2005 Lecture Notes in Computer Science  
Peer-to-peer networks have often been touted as the ultimate solution to scalability. Although cooperative techniques have been initially used almost exclusively for content lookup and sharing, one of the most promising application of the peer-to-peer paradigm is to capitalize the bandwidth of client peers to quickly distribute large content and withstand flash-crowds (i.e., a sudden increase in popularity of some online content). Cooperative content distribution is based on the premise that
more » ... the premise that the capacity of a network is as high as the sum of the resources of its nodes: the more peers in the network, the higher its aggregate bandwidth, and the better it can scale and serve new peers. Such networks can thus spontaneously adapt to the demand by taking advantage of available resources. In this paper, we evaluate the use of peer-to-peer networks for content distribution under various system assumptions, such as peer arrival rates, bandwidth capacities, cooperation strategies, or peer lifetimes. We argue that the self-scaling and self-organizing properties of cooperative networks pave the way for cost-effective, yet highly efficient and robust content distribution.
doi:10.1007/11428589_22 fatcat:bjo4l47qczf45kjf3zp2u2ki3a

Adaptive Load Balancing for DHT Lookups

Silvia Bianchi, Sabina Serbu, Pascal Felber, Peter Kropf
2006 Computer Communications and Networks  
In the past few years, several DHT-based abstractions for peer-to-peer systems have been proposed. The main characteristic is to associate nodes (peers) with objects (keys) and to construct distributed routing structures to support efficient location. These approaches partially consider the load problem by balancing storage of objects without, however, considering lookup traffic. In this paper we present an analysis of structured peer-to-peer systems taking into consideration Zipf-like requests
more » ... Zipf-like requests distribution. Based on our analysis, we propose a novel approach for load balancing taking into account object popularity. It is based on dynamic routing table reorganization in order to balance the routing load and on caching objects to balance the request load. We can therefore significantly improve the load balancing of traffic in these systems, and consequently their scalability and performance. Results from experimental evaluation demonstrate the effectiveness of our approach.
doi:10.1109/icccn.2006.286311 dblp:conf/icccn/BianchiSFK06 fatcat:osn4ihr75fectpjufwtqonuvea

A Lazy Snapshot Algorithm with Eager Validation [chapter]

Torvald Riegel, Pascal Felber, Christof Fetzer
2006 Lecture Notes in Computer Science  
Most high-performance software transactional memories (STM) use optimistic invisible reads. Consequently, a transaction might have an inconsistent view of the objects it accesses unless the consistency of the view is validated whenever the view changes. Although all STMs usually detect inconsistencies at commit time, a transaction might never reach this point because an inconsistent view can provoke arbitrary behavior in the application (e.g., enter an infinite loop). In this paper, we formally
more » ... paper, we formally introduce a lazy snapshot algorithm that verifies at each object access that the view observed by a transaction is consistent. Validating previously accessed objects is not necessary for that, however, it can be used on-demand to prolong the view's validity. We demonstrate both formally and by measurements that the performance of our approach is quite competitive by comparing other STMs with an STM that uses our algorithm.
doi:10.1007/11864219_20 fatcat:scgddykch5ek5ojksy64ijnbyi

ON THE INPUT ACCEPTANCE OF TRANSACTIONAL MEMORY

Vincent Gramoli, Derin Harmanci, Pascal Felber
2010 Parallel Processing Letters  
We present the Input Acceptance of Transactional Memory (TM). Despite the large interest for performance of TMs, no existing research work has investigated the impact of solving a conflict that does not need to be solved. Traditional solutions for a TM to be correct is to delay or abort a transaction as soon as it presents a risk to violate consistency. Both alternatives are costly and should be avoided if consistency is actually preserved. To address this problem, we introduce the input
more » ... ce the input acceptance of a TM as its ability to commit transactions, we upper-bound the input acceptance of existing TMs and propose a new TM with higher input acceptance.
doi:10.1142/s0129626410000041 fatcat:xs6bjdqd2jg5jc6wmgrxyufqq4

Twine: An Embedded Trusted Runtime for WebAssembly [article]

Jämes Ménétrey, Marcelo Pasin, Pascal Felber, Valerio Schiavoni
2021 arXiv   pre-print
WebAssembly is an increasingly popular lightweight binary instruction format, which can be efficiently embedded and sandboxed. Languages like C, C++, Rust, Go, and many others can be compiled into WebAssembly. This paper describes Twine, a WebAssembly trusted runtime designed to execute unmodified, language-independent applications. We leverage Intel SGX to build the runtime environment without dealing with language-specific, complex APIs. While SGX hardware provides secure execution within the
more » ... xecution within the processor, Twine provides a secure, sandboxed software runtime nested within an SGX enclave, featuring a WebAssembly system interface (WASI) for compatibility with unmodified WebAssembly applications. We evaluate Twine with a large set of general-purpose benchmarks and real-world applications. In particular, we used Twine to implement a secure, trusted version of SQLite, a well-known full-fledged embeddable database. We believe that such a trusted database would be a reasonable component to build many larger application services. Our evaluation shows that SQLite can be fully executed inside an SGX enclave via WebAssembly and existing system interface, with similar average performance overheads. We estimate that the performance penalties measured are largely compensated by the additional security guarantees and its full compatibility with standard WebAssembly. An in-depth analysis of our results indicates that performance can be greatly improved by modifying some of the underlying libraries. We describe and implement one such modification in the paper, showing up to 4.1× speedup. Twine is open-source, available at GitHub along with instructions to reproduce our experiments.
arXiv:2103.15860v1 fatcat:4k2flfy6qbab7lw3ahxkdbl7ki
« Previous Showing results 1 — 15 out of 400 results