The Experts below are selected from a list of 1617 Experts worldwide ranked by ideXlab platform
L. Lilien - One of the best experts on this subject based on the ideXlab platform.
-
ICDE - Quasi-partitioning: a new paradigm for Transaction Execution in partitioned distributed database systems
[1989] Proceedings. Fifth International Conference on Data Engineering, 1Co-Authors: L. LilienAbstract:The quasi-partitioning paradigm of operation for partitioned database systems is discussed in which a broken main link between two partitions can be replaced by a much slower backup link (e.g. a dial-up telephone connection). The paradigm solves the problem of preparation for network partitioning. The quasi-partitioning mode of operation has two primitive operations: creeping retrieval and creeping merge. Creeping retrieval increases data availability by crossing partition boundaries (over backup links) to read foreign data. Similarly, creeping merge improves the degree of partition-consistency by crossing partition boundaries to perform merge actions. A quasi-partitioning protocol consists of an adaptation protocol and a merge protocol. Taxonomies are shown for quasi-partitioning adaptation protocols and for quasi-partitioning merge protocols (for restoring partition-consistency after system reconnection). Since merge protocols and adaptation protocols are interdependent, it is indicated here how these protocols should be paired. >
Jeffrey Draper - One of the best experts on this subject based on the ideXlab platform.
-
A Filtering Mechanism to Reduce Network Bandwidth Utilization of Transaction Execution
ACM Transactions on Architecture and Code Optimization, 2016Co-Authors: Lihang Zhao, Lizhong Chen, Woojin Choi, Jeffrey DraperAbstract:Hardware Transactional Memory (HTM) relies heavily on the on-chip network for interTransaction communication. However, the network bandwidth utilization of Transactions has been largely neglected in HTM designs. In this work, we propose a cost model to analyze network bandwidth in Transaction Execution. The cost model identifies a set of key factors that can be optimized through system design to reduce the communication cost of HTM. Based on the model and network traffic characterization of a representative HTM design, we identify a huge source of superfluous traffic due to failed requests in Transaction conflicts. As observed in a spectrum of workloads, 39p of the Transactional requests fail due to conflicts, which renders 58p of the Transactional network traffic futile. To combat this pathology, a novel in-network filtering mechanism is proposed. The on-chip router is augmented to predict conflicts among Transactions and proactively filter out those requests that have a high probability to fail. Experimental results show the proposed mechanism reduces total network traffic by 24p on average for a set of high-contention TM applications, thereby reducing energy consumption by an average of 24p. Meanwhile, the contention in the coherence directory is reduced by 68p, on average. These improvements are achieved with only 5p area added to a conventional on-chip router design.
-
IPDPS - Mitigating the Mismatch between the Coherence Protocol and Conflict Detection in Hardware Transactional Memory
2014 IEEE 28th International Parallel and Distributed Processing Symposium, 2014Co-Authors: Lihang Zhao, Lizhong Chen, Jeffrey DraperAbstract:Hardware Transactional Memory (HTM) usually piggybacks onto the cache coherence protocol to detect data access conflicts between Transactions. We identify an intrinsic mismatch between the typical coherence scheme and Transaction Execution, which causes a sizable amount of unnecessary Transaction aborts. This pathological behavior is called false aborting and increases the amount of wasted computation and on-chip communication. For the TM applications we studied, 41% of the Transactional write requests incur false aborting. To combat false aborting, we propose Predictive Unicast and Notification (PUNO), a novel hardware mechanism to 1) replace the inefficient coherence multicast with a unicast scheme to prevent Transactions from being disrupted unnecessarily and 2) restrain Transaction polling through proactive notification. PUNO reduces Transaction aborts by 61% and network traffic by 32% in workloads representative of future TM applications with a VLSI implementation area overhead of 0.41%.
Roberto Palmieri - One of the best experts on this subject based on the ideXlab platform.
-
PPoPP - Processing Transactions in a predefined order
Proceedings of the 24th Symposium on Principles and Practice of Parallel Programming, 2019Co-Authors: Mohamed M Saad, Masoomeh Javidi Kishi, Shihao Jing, Sandeep Hans, Roberto PalmieriAbstract:In this paper we provide a high performance solution to the problem of committing Transactions while enforcing a pre-defined order. We provide the design and implementation of three algorithms, which deploy a specialized cooperative Transaction Execution model. This model permits the propagation of written values along the chain of ordered Transactions. We show that, even in the presence of data conflicts, the proposed algorithms outperform single threaded Execution, and other baseline and specialized state-of-the-art competitors (e.g., STMLite). The maximum speedup achieved in micro benchmarks, STAMP, PARSEC and SPEC200 applications is in the range of 4.3x -- 16.5x.
-
Processing Transactions in a Predefined Order
arXiv: Distributed Parallel and Cluster Computing, 2018Co-Authors: Mohamed M Saad, Masoomeh Javidi Kishi, Shihao Jing, Sandeep Hans, Roberto PalmieriAbstract:In this paper we provide a high performance solution to the problem of committing Transactions while enforcing a predefined order. We provide the design and implementation of three algorithms, which deploy a specialized cooperative Transaction Execution model. This model permits the propagation of written values along the chain of ordered Transactions. We show that, even in the presence of data conflicts, the proposed algorithms are able to outperform single-threaded Execution, and other baseline and specialized state-of-the-art competitors (e.g., STMLite). The maximum speedup achieved in micro benchmarks, STAMP, PARSEC and SPEC200 applications is in the range of 4.3x -- 16.5x.
-
PPOPP - On ordering Transaction commit
Proceedings of the 21st ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming - PPoPP '16, 2016Co-Authors: Mohamed M Saad, Roberto Palmieri, Binoy RavindranAbstract:In this poster paper, we briefly introduce an effective solution to address the problem of committing Transactions enforcing a predefined order. To do that, we overview the design of two algorithms that deploy a cooperative Transaction Execution that circumvents the Transaction isolation constraint in favor of propagating written values among conflicting Transactions. A preliminary implementation shows that even in the presence of data conflicts, the proposed algorithms outperform other competitors, significantly.
-
Transaction Execution models in partially replicated Transactional memory the case for data flow and control flow
Transactional Memory, 2015Co-Authors: Roberto Palmieri, Sebastiano Peluso, Binoy RavindranAbstract:In this chapter we describe solutions for managing concurrency of distributed Transactional memory accesses in partially replicated deployments. A system is classified as partially replicated if, for each shared object, there is more than one node responsible for storing the object, thus resulting in multiple copies available in the system. In contrast to full replication, where all objects are replicated on all nodes, partial replication allows storing a huge amount of data that, by nature, cannot fit in a single node and improving scalability by (significantly) increasing the number of node serving Transaction requests. Solutions that assume partially replicated deployments are categorized according to the mobility of shared objects. In the control-flow approach shared objects are pinned to nodes for the entire system’s lifetime, whereas in the data-flow objects are allowed to change residence node (also called owner) whenever a Transaction commits a new version of the object. Intuitively, adopting the data-flow model, objects follow committing Transactions whereas, relying on the control-flow model, Transactions’ flow is routed towards objects. There is a number of key factors to be evaluated before preferring one Transaction Execution model to another. This chapter surveys all of them and provides solutions suited for different deployments. The chapter aims for helping designers to understand the Execution model that better fits their requirements.
-
Transactional Memory - Transaction Execution Models in Partially Replicated Transactional Memory: The Case for Data-Flow and Control-Flow
Transactional Memory. Foundations Algorithms Tools and Applications, 2015Co-Authors: Roberto Palmieri, Sebastiano Peluso, Binoy RavindranAbstract:In this chapter we describe solutions for managing concurrency of distributed Transactional memory accesses in partially replicated deployments. A system is classified as partially replicated if, for each shared object, there is more than one node responsible for storing the object, thus resulting in multiple copies available in the system. In contrast to full replication, where all objects are replicated on all nodes, partial replication allows storing a huge amount of data that, by nature, cannot fit in a single node and improving scalability by (significantly) increasing the number of node serving Transaction requests. Solutions that assume partially replicated deployments are categorized according to the mobility of shared objects. In the control-flow approach shared objects are pinned to nodes for the entire system’s lifetime, whereas in the data-flow objects are allowed to change residence node (also called owner) whenever a Transaction commits a new version of the object. Intuitively, adopting the data-flow model, objects follow committing Transactions whereas, relying on the control-flow model, Transactions’ flow is routed towards objects. There is a number of key factors to be evaluated before preferring one Transaction Execution model to another. This chapter surveys all of them and provides solutions suited for different deployments. The chapter aims for helping designers to understand the Execution model that better fits their requirements.
Binoy Ravindran - One of the best experts on this subject based on the ideXlab platform.
-
PPOPP - On ordering Transaction commit
Proceedings of the 21st ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming - PPoPP '16, 2016Co-Authors: Mohamed M Saad, Roberto Palmieri, Binoy RavindranAbstract:In this poster paper, we briefly introduce an effective solution to address the problem of committing Transactions enforcing a predefined order. To do that, we overview the design of two algorithms that deploy a cooperative Transaction Execution that circumvents the Transaction isolation constraint in favor of propagating written values among conflicting Transactions. A preliminary implementation shows that even in the presence of data conflicts, the proposed algorithms outperform other competitors, significantly.
-
Transaction Execution models in partially replicated Transactional memory the case for data flow and control flow
Transactional Memory, 2015Co-Authors: Roberto Palmieri, Sebastiano Peluso, Binoy RavindranAbstract:In this chapter we describe solutions for managing concurrency of distributed Transactional memory accesses in partially replicated deployments. A system is classified as partially replicated if, for each shared object, there is more than one node responsible for storing the object, thus resulting in multiple copies available in the system. In contrast to full replication, where all objects are replicated on all nodes, partial replication allows storing a huge amount of data that, by nature, cannot fit in a single node and improving scalability by (significantly) increasing the number of node serving Transaction requests. Solutions that assume partially replicated deployments are categorized according to the mobility of shared objects. In the control-flow approach shared objects are pinned to nodes for the entire system’s lifetime, whereas in the data-flow objects are allowed to change residence node (also called owner) whenever a Transaction commits a new version of the object. Intuitively, adopting the data-flow model, objects follow committing Transactions whereas, relying on the control-flow model, Transactions’ flow is routed towards objects. There is a number of key factors to be evaluated before preferring one Transaction Execution model to another. This chapter surveys all of them and provides solutions suited for different deployments. The chapter aims for helping designers to understand the Execution model that better fits their requirements.
-
Transactional Memory - Transaction Execution Models in Partially Replicated Transactional Memory: The Case for Data-Flow and Control-Flow
Transactional Memory. Foundations Algorithms Tools and Applications, 2015Co-Authors: Roberto Palmieri, Sebastiano Peluso, Binoy RavindranAbstract:In this chapter we describe solutions for managing concurrency of distributed Transactional memory accesses in partially replicated deployments. A system is classified as partially replicated if, for each shared object, there is more than one node responsible for storing the object, thus resulting in multiple copies available in the system. In contrast to full replication, where all objects are replicated on all nodes, partial replication allows storing a huge amount of data that, by nature, cannot fit in a single node and improving scalability by (significantly) increasing the number of node serving Transaction requests. Solutions that assume partially replicated deployments are categorized according to the mobility of shared objects. In the control-flow approach shared objects are pinned to nodes for the entire system’s lifetime, whereas in the data-flow objects are allowed to change residence node (also called owner) whenever a Transaction commits a new version of the object. Intuitively, adopting the data-flow model, objects follow committing Transactions whereas, relying on the control-flow model, Transactions’ flow is routed towards objects. There is a number of key factors to be evaluated before preferring one Transaction Execution model to another. This chapter surveys all of them and provides solutions suited for different deployments. The chapter aims for helping designers to understand the Execution model that better fits their requirements.
Lihang Zhao - One of the best experts on this subject based on the ideXlab platform.
-
A Filtering Mechanism to Reduce Network Bandwidth Utilization of Transaction Execution
ACM Transactions on Architecture and Code Optimization, 2016Co-Authors: Lihang Zhao, Lizhong Chen, Woojin Choi, Jeffrey DraperAbstract:Hardware Transactional Memory (HTM) relies heavily on the on-chip network for interTransaction communication. However, the network bandwidth utilization of Transactions has been largely neglected in HTM designs. In this work, we propose a cost model to analyze network bandwidth in Transaction Execution. The cost model identifies a set of key factors that can be optimized through system design to reduce the communication cost of HTM. Based on the model and network traffic characterization of a representative HTM design, we identify a huge source of superfluous traffic due to failed requests in Transaction conflicts. As observed in a spectrum of workloads, 39p of the Transactional requests fail due to conflicts, which renders 58p of the Transactional network traffic futile. To combat this pathology, a novel in-network filtering mechanism is proposed. The on-chip router is augmented to predict conflicts among Transactions and proactively filter out those requests that have a high probability to fail. Experimental results show the proposed mechanism reduces total network traffic by 24p on average for a set of high-contention TM applications, thereby reducing energy consumption by an average of 24p. Meanwhile, the contention in the coherence directory is reduced by 68p, on average. These improvements are achieved with only 5p area added to a conventional on-chip router design.
-
IPDPS - Mitigating the Mismatch between the Coherence Protocol and Conflict Detection in Hardware Transactional Memory
2014 IEEE 28th International Parallel and Distributed Processing Symposium, 2014Co-Authors: Lihang Zhao, Lizhong Chen, Jeffrey DraperAbstract:Hardware Transactional Memory (HTM) usually piggybacks onto the cache coherence protocol to detect data access conflicts between Transactions. We identify an intrinsic mismatch between the typical coherence scheme and Transaction Execution, which causes a sizable amount of unnecessary Transaction aborts. This pathological behavior is called false aborting and increases the amount of wasted computation and on-chip communication. For the TM applications we studied, 41% of the Transactional write requests incur false aborting. To combat false aborting, we propose Predictive Unicast and Notification (PUNO), a novel hardware mechanism to 1) replace the inefficient coherence multicast with a unicast scheme to prevent Transactions from being disrupted unnecessarily and 2) restrain Transaction polling through proactive notification. PUNO reduces Transaction aborts by 61% and network traffic by 32% in workloads representative of future TM applications with a VLSI implementation area overhead of 0.41%.
