An optimistic checkpointing and selective message logging approach for consistent global checkpoint collection in distributed systems

Jiang, Qiangfeng; Manivannan, D.

doi:10.1109/ipdps.2007.370308

Cited by 23 publications

(10 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Another approach that has been suggested to improve the performance of checkpointing systems is uncoordinated or asynchronous checkpointing [2], [8], [9]. In these systems, nodes generally checkpoint and restore from local storage without the synchronization used by coordinated checkpointing.…”

Section: B Uncoordinated Checkpointingmentioning

confidence: 99%

Energy Consumption of Resilience Mechanisms in Large Scale Systems

Mills

Znati

Melhem

et al. 2014

2014 22nd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing

View full text Add to dashboard Cite

As HPC systems continue to grow to meet the requirements of tomorrow's exascale-class systems, two of the biggest challenges are power consumption and system resilience. On current systems, the dominant resilience technique is checkpoint/restart. It is believed, however, that this technique alone will not scale to the level necessary to support future systems. Therefore, alternative methods have been suggested to augment checkpoint/restart -for example process replication. In this paper we address both resilience and power together, this is in contrast to much of the competed work which does so independently. Using an analytical model that accounts for both power consumption and failures, we study the performance of checkpoint and replication-based techniques on current and future systems and use power measurements from current systems to validate our findings. Lastly, in an attempt to optimize power consumption for replication, we introduce a new protocol termed shadow replication which not only reduces energy consumption but also produces faster response times than checkpoint/restart and traditional replication when operating under system power constraints.

show abstract

Section: B Uncoordinated Checkpointingmentioning

confidence: 99%

Energy Consumption of Resilience Mechanisms in Large Scale Systems

Mills

Znati

Melhem

et al. 2014

2014 22nd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing

View full text Add to dashboard Cite

show abstract

“…The protocol does not force every node to take a local snapshot every time a global checkpoint is taken and also does not block the computation during snapshot collection. A two phase asynchronous or uncoordinated checkpointing scheme for mobile devices is presented in [8]. This scheme guarantees that every checkpoint taken by a process will be a part of a consistent global checkpoint.…”

Section: Related Workmentioning

confidence: 99%

“…A large number of checkpointing based rollback recovery protocols have been designed for cellular mobile computing systems in the past [6][7][8]. However, these algorithms assume the presence of fixed infrastructure in the form of static hosts or Mobile Support Stations for providing sufficient stable storage (for saving the checkpoints and message logs) and handling the mobility of hosts.…”

Section: Introductionmentioning

confidence: 99%

Adaptive checkpointing for fault tolerance in an autonomous mobile computing grid

Jaggi

Singh

2014

2014 International Conference on Contemporary Computing and Informatics (IC3I)

View full text Add to dashboard Cite

The widespread availability and increasing processing power of mobile devices has lead to a focus towards the development of autonomous mobile computing grids (MoGs). Such mobile grids allow the successful execution of distributed applications; without access to any static nodes or wired networks. However, the implementation of a fault tolerance technique is essential to completely utilize the mobile devices as viable computing resources. Checkpointing is a well explored fault tolerance technique for mobile computing systems. The paper presents an adaptive checkpointing technique for failure recovery of mobile nodes in a MoG. The presented protocol relies on cooperative checkpointing by the constituent nodes in the system. A node uses the stable storage of other nodes in the system to save its checkpoint data, in case the requisite stable storage is not available at the node itself. Further, depending on the availability of resources in the MoG, the scheme replicates a node's checkpoint data at multiple nodes. The results of simulation for the presented scheme verify that the introduction of redundancy with the checkpointing procedure vastly increases the likelihood of successful recovery of a failed node in a MoG.

show abstract

“…Another approach that has been suggested to improve the performance of checkpointing systems is uncoordinated or asynchronous checkpointing [1,23,24]. These methods typically checkpoint and restore from local storage without the synchronization used by coordinated checkpointing.…”

Section: Asynchronous Checkpointing and Message Loggingmentioning

confidence: 99%