José Duato
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
José Duato has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
- Pedro Lopez
Universitat Politécnica de Valencia, Valencia, Spain
5 shared publications - Francisco Jose Quiles
Universidad de Castilla-La Mancha Altagracia 50 Ciudad Real 13071 Ciudad Real Spain
3 shared publications - Juan Manuel Orduña
Departamento de Informática, Universidad de Valencia, Burjassot, Spain
3 shared publications - Rafael Casado
University of Castilla-La Mancha
3 shared publications - Silvia Rueda
Universidad de Valencia
3 shared publications J. Flich
Dept. of Comput. Eng., Univ. Politecnica de Valencia, Spain
2 shared publications
Latest Publications
- 29 July 2011
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Parallel and Distributed Systems
- Vol. 22 (10) , 1610-1623
Efficient and Scalable Starvation Prevention Mechanism for Token Coherence
Token Coherence is a cache coherence protocol that simultaneously captures the best attributes of the traditional approximations to coherence: direct communication between processors (like snooping-based protocols) and no reliance on bus-like interconnects (like directory-based protocols). This is possible thanks to a class of unordered requests that usually succeed in resolving the cache misses. The problem of the unordered requests is that they can cause protocol races, which prevent some misses from being resolved. To eliminate races and ensure the completion of the unresolved misses, Token Coherence uses a starvation prevention mechanism named persistent requests. This mechanism is extremely inefficient and, besides, it endangers the scalability of Token Coherence since it requires storage structures (at each node) whose size grows proportionally to the system size. While multiprocessors continue including an increasingly number of nodes, both the performance and scalability of cache coherence protocols will continue to be key aspects. In this work, we propose an alternative starvation prevention mechanism, named priority requests, that outperforms the persistent request one. This mechanism is able to reduce the application runtime more than 20 percent (on average) in a 64-processor system. Furthermore, thanks to the flexibility shown by priority requests, it is possible to drastically minimize its storage requirements, thereby improving the whole scalability of Token Coherence. Although this is achieved at the expense of a slight performance degradation, priority requests still outperform persistent requests significantly.
- 29 July 2011
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Parallel and Distributed Systems
- Vol. 22 (10) , 1610-1623
Efficient and Scalable Starvation Prevention Mechanism for Token Coherence
Token Coherence is a cache coherence protocol that simultaneously captures the best attributes of the traditional approximations to coherence: direct communication between processors (like snooping-based protocols) and no reliance on bus-like interconnects (like directory-based protocols). This is possible thanks to a class of unordered requests that usually succeed in resolving the cache misses. The problem of the unordered requests is that they can cause protocol races, which prevent some misses from being resolved. To eliminate races and ensure the completion of the unresolved misses, Token Coherence uses a starvation prevention mechanism named persistent requests. This mechanism is extremely inefficient and, besides, it endangers the scalability of Token Coherence since it requires storage structures (at each node) whose size grows proportionally to the system size. While multiprocessors continue including an increasingly number of nodes, both the performance and scalability of cache coherence protocols will continue to be key aspects. In this work, we propose an alternative starvation prevention mechanism, named priority requests, that outperforms the persistent request one. This mechanism is able to reduce the application runtime more than 20 percent (on average) in a 64-processor system. Furthermore, thanks to the flexibility shown by priority requests, it is possible to drastically minimize its storage requirements, thereby improving the whole scalability of Token Coherence. Although this is achieved at the expense of a slight performance degradation, priority requests still outperform persistent requests significantly.
- 14 April 2010
- Published by Elsevier in Future Generation Computer Systems
- Vol. 26 (7) , 905-915
Ensuring the performance and scalability of peer-to-peer distributed virtual environments
- 14 April 2010
- Published by Elsevier in Future Generation Computer Systems
- Vol. 26 (7) , 905-915
Ensuring the performance and scalability of peer-to-peer distributed virtual environments
- 31 July 2008
- Published by Elsevier in Journal of Systems Architecture
- Vol. 54 (7) , 664-678
A proposal for managing ASI fabrics
- 31 July 2008
- Published by Elsevier in Journal of Systems Architecture
- Vol. 54 (7) , 664-678
A proposal for managing ASI fabrics
- 1 March 2007
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 10878270,p. 107-114
On the Characterization of Peer-To-Peer Distributed Virtual Environments
Large scale distributed virtual environments (DVEs) have become a major trend in distributed applications, mainly due to the enormous popularity of multi-player online games in the entertainment industry. Since architectures based on networked servers seem to be not scalable enough to support massively multi-player applications, peer-to-peer (P2P) architectures have been proposed as an efficient and truly scalable solution for this kind of systems. However, in order to design efficient DVEs based on peer-to-peer architectures these systems must be characterized, measuring the impact of different client behaviors on system performance. This paper presents the experimental characterization of peer-to-peer distributed virtual environments in regard to well-known performance metrics in distributed systems. Characterization results show that system saturation is inherently avoided due to the peer-to-peer scheme, as it could be expected. Also, these results show that the saturation of a given client exclusively has an effect on the surrounding clients in the virtual world, having no noticeable effect at all on the rest of avatars. Finally, the characterization results show that the response time offered to client computers greatly depends on the number of new connections that these clients have to make when new neighbors appear in the virtual world. These results can be used as the basis for an efficient design of peer-to-peer DVE systems.
- 1 March 2007
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 10878270,p. 107-114
On the Characterization of Peer-To-Peer Distributed Virtual Environments
Large scale distributed virtual environments (DVEs) have become a major trend in distributed applications, mainly due to the enormous popularity of multi-player online games in the entertainment industry. Since architectures based on networked servers seem to be not scalable enough to support massively multi-player applications, peer-to-peer (P2P) architectures have been proposed as an efficient and truly scalable solution for this kind of systems. However, in order to design efficient DVEs based on peer-to-peer architectures these systems must be characterized, measuring the impact of different client behaviors on system performance. This paper presents the experimental characterization of peer-to-peer distributed virtual environments in regard to well-known performance metrics in distributed systems. Characterization results show that system saturation is inherently avoided due to the peer-to-peer scheme, as it could be expected. Also, these results show that the saturation of a given client exclusively has an effect on the surrounding clients in the virtual world, having no noticeable effect at all on the rest of avatars. Finally, the characterization results show that the response time offered to client computers greatly depends on the number of new connections that these clients have to make when new neighbors appear in the virtual world. These results can be used as the basis for an efficient design of peer-to-peer DVE systems.
- 1 February 2007
- Published by Elsevier in Computer Communications
- Vol. 30 (4) , 731-739
A genetic approach for adding QoS to distributed virtual environments
- 1 February 2007
- Published by Elsevier in Computer Communications
- Vol. 30 (4) , 731-739
A genetic approach for adding QoS to distributed virtual environments
- 22 March 2004
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 10 pp.
VOQ/sub SW/: a methodology to reduce HOL blocking in InfiniBand networks
InfiniBand is a new switch-based standard interconnect for communication between processor nodes and I/O devices as well as for interprocessor communication. InfiniBand architecture allows switches to support up to 15 virtual lanes per port for data traffic. To route packets through a given virtual lane (VL), packets are labeled with a certain service level (SL) at injection time, and SLtoVL mapping tables are used at each switch to determine the VL to be used. Many previous works in the literature have shown that separate virtual lanes are able to reduce the influence of the well-known head-of-line (HOL) blocking effect on network performance. However, using virtual lanes to form separate virtual networks is not enough to eliminate the HOL blocking problem. Alternative solutions such as Virtual Output Queuing (VOQ) are able to eliminate it at the expense of modifying the switch buffer organization. In this paper, we propose an effective strategy to implement the VOQ scheme in IBA switches by using virtual lanes. This strategy does not require to modify the switch architecture, simply SL to VL tables must be properly filled. Evaluation results show that our proposed VOQ scheme is able to outperform the results obtained with the virtual network approach using the same number of resources. Moreover, the methodology proposed to implement the VOQ scheme in IBA only requires a small number of resources in order to significantly improve network throughput.
- 22 March 2004
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 10 pp.
VOQ/sub SW/: a methodology to reduce HOL blocking in InfiniBand networks
InfiniBand is a new switch-based standard interconnect for communication between processor nodes and I/O devices as well as for interprocessor communication. InfiniBand architecture allows switches to support up to 15 virtual lanes per port for data traffic. To route packets through a given virtual lane (VL), packets are labeled with a certain service level (SL) at injection time, and SLtoVL mapping tables are used at each switch to determine the VL to be used. Many previous works in the literature have shown that separate virtual lanes are able to reduce the influence of the well-known head-of-line (HOL) blocking effect on network performance. However, using virtual lanes to form separate virtual networks is not enough to eliminate the HOL blocking problem. Alternative solutions such as Virtual Output Queuing (VOQ) are able to eliminate it at the expense of modifying the switch buffer organization. In this paper, we propose an effective strategy to implement the VOQ scheme in IBA switches by using virtual lanes. This strategy does not require to modify the switch architecture, simply SL to VL tables must be properly filled. Evaluation results show that our proposed VOQ scheme is able to outperform the results obtained with the virtual network approach using the same number of resources. Moreover, the methodology proposed to implement the VOQ scheme in IBA only requires a small number of resources in order to significantly improve network throughput.
- 1 January 2004
- Published by Springer Nature
- p. 93-107
Collective Communication Patterns on the Quadrics Network
- 1 January 2004
- Published by Springer Nature
- p. 93-107
Collective Communication Patterns on the Quadrics Network
- 25 June 2003
- Published by Institute of Electrical and Electronics Engineers (IEEE)
Removing the latency overhead of the ITB mechanism in COWs with source routing
Clusters of workstations (COWs) are becoming increasingly popular as a cost-effective alternative to parallel computers. The in-transit buffer (ITB) mechanism can improve network performance when applied to COWs with irregular topology and source routing. This mechanism considerably improves the performance of this kind of network when compared to current source routing algorithms; however, it introduces a latency penalty. An implementation of this mechanism was performed, showing that the latency overhead of the mechanism may be noticeable, especially for short messages and at low network loads. In this paper, we analyze in detail the latency overhead of ITBs, proposing several mechanisms to reduce, hide and remove it. Firstly, we show, by simulation, the effect of an ITB implementation that is much slower than the one implemented. Then we propose three mechanisms that try to overcome the latency penalty. All the mechanisms are simple and can be easily implemented; also, they are out of the critical path of the ITB packet-processing procedure. The results show very good behaviour of the proposed mechanisms, considerably reducing or even completely removing the latency overhead.
- 25 June 2003
- Published by Institute of Electrical and Electronics Engineers (IEEE)
Removing the latency overhead of the ITB mechanism in COWs with source routing
Clusters of workstations (COWs) are becoming increasingly popular as a cost-effective alternative to parallel computers. The in-transit buffer (ITB) mechanism can improve network performance when applied to COWs with irregular topology and source routing. This mechanism considerably improves the performance of this kind of network when compared to current source routing algorithms; however, it introduces a latency penalty. An implementation of this mechanism was performed, showing that the latency overhead of the mechanism may be noticeable, especially for short messages and at low network loads. In this paper, we analyze in detail the latency overhead of ITBs, proposing several mechanisms to reduce, hide and remove it. Firstly, we show, by simulation, the effect of an ITB implementation that is much slower than the one implemented. Then we propose three mechanisms that try to overcome the latency penalty. All the mechanisms are simple and can be easily implemented; also, they are out of the critical path of the ITB packet-processing procedure. The results show very good behaviour of the proposed mechanisms, considerably reducing or even completely removing the latency overhead.
- 13 November 2002
- Published by Institute of Electrical and Electronics Engineers (IEEE)
A congestion control mechanism for wormhole networks
- 13 November 2002
- Published by Institute of Electrical and Electronics Engineers (IEEE)
A congestion control mechanism for wormhole networks
- 1 January 2000
- Published by Springer Nature
- p. 207-231
Dynamic Reconfiguration in High Speed Local Area Networks
- 1 January 2000
- Published by Springer Nature
- p. 207-231
Dynamic Reconfiguration in High Speed Local Area Networks