The Experts below are selected from a list of 11607 Experts worldwide ranked by ideXlab platform
Mike Dahlin - One of the best experts on this subject based on the ideXlab platform.
-
gnothi separating data and metadata for efficient and available Storage Replication
USENIX Annual Technical Conference, 2012Co-Authors: Yang Wang, Lorenzo Alvisi, Mike DahlinAbstract:This paper describes Gnothi, a block Replication system that separates data from metadata to provide efficient and available Storage Replication. Separating data from metadata allows Gnothi to execute disk accesses on subsets of replicas while using fully replicated metadata to ensure that requests are executed correctly and to speed up recovery of slow or failed replicas. Performance evaluation shows that Gnothi can achieve 40-64% higher write throughput than previous work and significantly save Storage space. Furthermore, while a failed replica recovers, Gnothi can provide about 100- 200% higher throughput, while still retaining the same recovery time and while guaranteeing that recovery eventually completes.
-
USENIX Annual Technical Conference - Gnothi: separating data and metadata for efficient and available Storage Replication
2012Co-Authors: Yang Wang, Lorenzo Alvisi, Mike DahlinAbstract:This paper describes Gnothi, a block Replication system that separates data from metadata to provide efficient and available Storage Replication. Separating data from metadata allows Gnothi to execute disk accesses on subsets of replicas while using fully replicated metadata to ensure that requests are executed correctly and to speed up recovery of slow or failed replicas. Performance evaluation shows that Gnothi can achieve 40-64% higher write throughput than previous work and significantly save Storage space. Furthermore, while a failed replica recovers, Gnothi can provide about 100- 200% higher throughput, while still retaining the same recovery time and while guaranteeing that recovery eventually completes.
M. Van Steen - One of the best experts on this subject based on the ideXlab platform.
-
a probabilistic Replication and Storage scheme for large wireless networks of small devices
Mobile Adhoc and Sensor Systems, 2008Co-Authors: Daniela Gavidia, M. Van SteenAbstract:Nodes in wireless ad hoc networks are often limited in terms of resources, such as Storage, power, and bandwidth. A downside of this is the fact that local Storage at one node cannot accommodate the vast amount of data contained in the network. In this paper, we present SharedState, a scheme for Storage, Replication, and distribution of common-interest data in wireless networks of resource-constrained devices (e.g. sensor nodes or embedded devices). SharedState works under the assumption that individual nodes would greatly benefit from having access to the wealth of information in the network, but are unable to store it locally at once. SharedState strives to make data available to every node by providing local access to a subset of the whole collection of data items in the network at any moment in time and ensuring that this subset is updated periodically. This is accomplished by probabilistic propagation and Replication of data items, ensuring the availability and persistence of information in the face of changing network conditions. We evaluate the performance of SharedState by studying the effectiveness with which nodes can gather information from the network. In addition, we optimize the bandwidth usage of our proposed solution by minimizing unnecessary communication based on feedback from the local neighborhood.
-
MASS - A probabilistic Replication and Storage scheme for large wireless networks of small devices
2008 5th IEEE International Conference on Mobile Ad Hoc and Sensor Systems, 2008Co-Authors: Daniela Gavidia, M. Van SteenAbstract:Nodes in wireless ad hoc networks are often limited in terms of resources, such as Storage, power, and bandwidth. A downside of this is the fact that local Storage at one node cannot accommodate the vast amount of data contained in the network. In this paper, we present SharedState, a scheme for Storage, Replication, and distribution of common-interest data in wireless networks of resource-constrained devices (e.g. sensor nodes or embedded devices). SharedState works under the assumption that individual nodes would greatly benefit from having access to the wealth of information in the network, but are unable to store it locally at once. SharedState strives to make data available to every node by providing local access to a subset of the whole collection of data items in the network at any moment in time and ensuring that this subset is updated periodically. This is accomplished by probabilistic propagation and Replication of data items, ensuring the availability and persistence of information in the face of changing network conditions. We evaluate the performance of SharedState by studying the effectiveness with which nodes can gather information from the network. In addition, we optimize the bandwidth usage of our proposed solution by minimizing unnecessary communication based on feedback from the local neighborhood.
Yang Wang - One of the best experts on this subject based on the ideXlab platform.
-
gnothi separating data and metadata for efficient and available Storage Replication
USENIX Annual Technical Conference, 2012Co-Authors: Yang Wang, Lorenzo Alvisi, Mike DahlinAbstract:This paper describes Gnothi, a block Replication system that separates data from metadata to provide efficient and available Storage Replication. Separating data from metadata allows Gnothi to execute disk accesses on subsets of replicas while using fully replicated metadata to ensure that requests are executed correctly and to speed up recovery of slow or failed replicas. Performance evaluation shows that Gnothi can achieve 40-64% higher write throughput than previous work and significantly save Storage space. Furthermore, while a failed replica recovers, Gnothi can provide about 100- 200% higher throughput, while still retaining the same recovery time and while guaranteeing that recovery eventually completes.
-
USENIX Annual Technical Conference - Gnothi: separating data and metadata for efficient and available Storage Replication
2012Co-Authors: Yang Wang, Lorenzo Alvisi, Mike DahlinAbstract:This paper describes Gnothi, a block Replication system that separates data from metadata to provide efficient and available Storage Replication. Separating data from metadata allows Gnothi to execute disk accesses on subsets of replicas while using fully replicated metadata to ensure that requests are executed correctly and to speed up recovery of slow or failed replicas. Performance evaluation shows that Gnothi can achieve 40-64% higher write throughput than previous work and significantly save Storage space. Furthermore, while a failed replica recovers, Gnothi can provide about 100- 200% higher throughput, while still retaining the same recovery time and while guaranteeing that recovery eventually completes.
Daniela Gavidia - One of the best experts on this subject based on the ideXlab platform.
-
a probabilistic Replication and Storage scheme for large wireless networks of small devices
Mobile Adhoc and Sensor Systems, 2008Co-Authors: Daniela Gavidia, M. Van SteenAbstract:Nodes in wireless ad hoc networks are often limited in terms of resources, such as Storage, power, and bandwidth. A downside of this is the fact that local Storage at one node cannot accommodate the vast amount of data contained in the network. In this paper, we present SharedState, a scheme for Storage, Replication, and distribution of common-interest data in wireless networks of resource-constrained devices (e.g. sensor nodes or embedded devices). SharedState works under the assumption that individual nodes would greatly benefit from having access to the wealth of information in the network, but are unable to store it locally at once. SharedState strives to make data available to every node by providing local access to a subset of the whole collection of data items in the network at any moment in time and ensuring that this subset is updated periodically. This is accomplished by probabilistic propagation and Replication of data items, ensuring the availability and persistence of information in the face of changing network conditions. We evaluate the performance of SharedState by studying the effectiveness with which nodes can gather information from the network. In addition, we optimize the bandwidth usage of our proposed solution by minimizing unnecessary communication based on feedback from the local neighborhood.
-
MASS - A probabilistic Replication and Storage scheme for large wireless networks of small devices
2008 5th IEEE International Conference on Mobile Ad Hoc and Sensor Systems, 2008Co-Authors: Daniela Gavidia, M. Van SteenAbstract:Nodes in wireless ad hoc networks are often limited in terms of resources, such as Storage, power, and bandwidth. A downside of this is the fact that local Storage at one node cannot accommodate the vast amount of data contained in the network. In this paper, we present SharedState, a scheme for Storage, Replication, and distribution of common-interest data in wireless networks of resource-constrained devices (e.g. sensor nodes or embedded devices). SharedState works under the assumption that individual nodes would greatly benefit from having access to the wealth of information in the network, but are unable to store it locally at once. SharedState strives to make data available to every node by providing local access to a subset of the whole collection of data items in the network at any moment in time and ensuring that this subset is updated periodically. This is accomplished by probabilistic propagation and Replication of data items, ensuring the availability and persistence of information in the face of changing network conditions. We evaluate the performance of SharedState by studying the effectiveness with which nodes can gather information from the network. In addition, we optimize the bandwidth usage of our proposed solution by minimizing unnecessary communication based on feedback from the local neighborhood.
Lorenzo Alvisi - One of the best experts on this subject based on the ideXlab platform.
-
gnothi separating data and metadata for efficient and available Storage Replication
USENIX Annual Technical Conference, 2012Co-Authors: Yang Wang, Lorenzo Alvisi, Mike DahlinAbstract:This paper describes Gnothi, a block Replication system that separates data from metadata to provide efficient and available Storage Replication. Separating data from metadata allows Gnothi to execute disk accesses on subsets of replicas while using fully replicated metadata to ensure that requests are executed correctly and to speed up recovery of slow or failed replicas. Performance evaluation shows that Gnothi can achieve 40-64% higher write throughput than previous work and significantly save Storage space. Furthermore, while a failed replica recovers, Gnothi can provide about 100- 200% higher throughput, while still retaining the same recovery time and while guaranteeing that recovery eventually completes.
-
USENIX Annual Technical Conference - Gnothi: separating data and metadata for efficient and available Storage Replication
2012Co-Authors: Yang Wang, Lorenzo Alvisi, Mike DahlinAbstract:This paper describes Gnothi, a block Replication system that separates data from metadata to provide efficient and available Storage Replication. Separating data from metadata allows Gnothi to execute disk accesses on subsets of replicas while using fully replicated metadata to ensure that requests are executed correctly and to speed up recovery of slow or failed replicas. Performance evaluation shows that Gnothi can achieve 40-64% higher write throughput than previous work and significantly save Storage space. Furthermore, while a failed replica recovers, Gnothi can provide about 100- 200% higher throughput, while still retaining the same recovery time and while guaranteeing that recovery eventually completes.