ZooKeeper: Wait-free Coordination for Internet-scale Systems

By Patrick Hunt, Mahadev Konar, Flavio Paiva Junqueira, Benjamin C. Reed. 2010

1,887 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In ML Systems & Hardware

Citations in the corpus

TensorFlow: A system for large-scale machine learning (2016, neutral): “Although we do not use strong consistency for the training state, we rely on a system like Chubby [8] or ZooKeeper [31] to map task IDs to IP addresses.”
Trends in big data analytics (2014, methodology): “The open-source equivalents of these systems, HDFS [44]2 and HBase [43], use Zookeeper [51] for consistency; Zookeeper uses the Zab protocol [78] to propagate the incremental state updates from primary node to backup nodes.”
RBFT: Redundant Byzantine Fault Tolerance (2013, background): “Note that RBFT is intended for open loop systems (such as e.g., Zookeeper [12] or Boxwood [16] asynchronous API), i.e., systems where a client may send multiple requests in parallel without waiting the reception of replies of anterior requests.”
Zab: High-performance broadcast for primary-backup systems (2011, background): “[1] ZooKeeper is a replicated service, and it requires that a majority (or more generally a quorum) of servers has not crashed for progress.”
Fault Tolerance in Iterative-Convergent Machine Learning (2018, methodology): “The detection of failures is performed by a failure detector service, which can leverage heartbeating mechanisms in existing systems for distributed consensus such as ZooKeeper (Hunt et al., 2010).”
SoK: Consensus in the Age of Blockchains (2017, methodology): “For example, Hyperledger supports a CFT service based on Apache Kafka [2] and its ZooKeeper unit [78], and more recently a PBFT-variant BFTSmart [30].”
Fast crash recovery in RAMCloud (2011, methodology): “Coordinator failures will be handled safely using the ZooKe ep r service [14].”
A Hitchhiker's Guide On Distributed Training of Deep Neural Networks (2018, methodology): “There have been several modifications proposed in the Raft protocol in systems like Zookeeper [32], Kafka [33] to name a few.”
No compromises: distributed transactions with consistency, availability, and performance (2015, background): “This illustrates the importance of reducing CPU overhead to realize the potential of the new hardware.”
Kineograph: taking the pulse of a fast-changing and connected world (2012, methodology): “ZooKeeper: Wait-free coordination for internet-scale systems.”
The RAMCloud Storage System (2015, methodology): “There are currently two implementations of this interface, one based on ZooKeeper [Hunt et al. 2010] and the other based on LogCabin, a storage system that uses the Raft consensus protocol [GitHub 2015b; Ongaro and Ousterhout 2014].”
The rise of "big data" on cloud computing: Review and open research issues (2015, background): “…Pig framework involves a high-level scripting language (Pig Latin) and offers a run-time platform that allows users to execute MapReduce on Hadoop [81] Zookeeper TM High-performance service to coordinate the processes of distributed applications; ZooKeeper allows distributed processes to manage…”
Split/Merge: System Support for Elastic Execution in Virtual Middleboxes (2013, neutral): “In the rare case that strong consistency is required, the application does not specify a combiner, and library instead interacts with a distributed locking service [3,11].”
ONOS: towards an open, distributed SDN OS (2014, methodology): “In our current design, network state is written to RAMCloud sequentially, so the throughput is simply the inverse of the latency. the system for updating network state in response to events; examples include rerouting traffic in response to link failure, moving traffic in response to congestion, and”
NetChain: Scale-Free Sub-RTT Coordination (2018, methodology): “Today’s server-based solutions require multiple endto-end round-trip times (RTTs) to process a query [1, 2, 3]: a client sends a request to coordination servers; the coordination servers execute a consensus protocol, which can take several RTTs; the coordination servers send a reply back to the clie”