Fast crash recovery in RAMCloud

By Diego Ongaro, Stephen M. Rumble, Ryan Stutsman, John K. Ousterhout, Mendel Rosenblum. Symposium on Operating Systems Principles, 2011

387 citations recorded by the source corpus.

Subfields: Computer Networks and Communications, Hardware and Architecture

View Paper: https://www.semanticscholar.org/paper/e3d76cf4e0718c8e00750c8704e2cb1712e409be

How researchers see this paper

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

In Computer Networks and Communications

In Hardware and Architecture

Citations in the corpus

No compromises: distributed transactions with consistency, availability, and performance (2015, background): “The number of false positives is reduced by using unreliable datagrams but it is not eliminated due to contention for the CPU.”
Zero-effort payments: design, deployment, and lessons (2014, neutral): “Fast crash recovery is a classic technique to improve availability in complex systems [3, 7, 18].”
Less Is More: Trading a Little Bandwidth for Ultra-Low Latency in the Data Center (2012, background): “Several applications and platforms have recently arisen that necessitate very low latency RPCs; for example, high-frequency trading (see [30]), high-performance computing, and RAMCloud [37, 38].”
IX: A Protected Dataplane Operating System for High Throughput and Low Latency (2014, methodology): “USENIX Association 11th USENIX Symposium on Operating Systems Design and Implementation (OSDI ’14) 49”
MemC3: Compact and Concurrent MemCache with Dumber Caching and Smarter Hashing (2013, neutral): “RAMCloud [22] focused on fast data reconstruction from on-disk replicas.”
In-Memory Big Data Management and Processing: A Survey (2015, background): “High availability is usually achieved by maintaining multiple replicas and stand-by servers [37], [148], [149], [150], or relying on fast recovery upon failure [49], [126].”
Discretized streams: fault-tolerant streaming computation at scale (2013, methodology): “Our parallel recovery mechanism is also similar to techniques in MapReduce, GFS, and RAMCloud [12, 16, 30], all of which partition recovery work on failure.”
Discretized Streams: An Efficient and Fault-Tolerant Model for Stream Processing on Large Clusters (2012, methodology): “Finally, our parallel recovery mechanism is conceptually similar to recovery techniques in MapReduce, GFS, and RAMCloud [6, 9, 17], which all leverage repartitioning.”
Trinity: a distributed graph engine on a memory cloud (2013, background): “Detailed discussion about buffered logging can be found in [30, 29].”
Datacenter RPCs can be General and Fast (2018, background): “Absent networking options that provide both high performance and generality, recent systems often choose to design and implement their own communication layer using low-level interfaces [49, 36, 24, 25, 52, 61, 17, 37].”
FaRM: Fast Remote Memory (2014, neutral): “A cluster with one hundred machines can hold tens of terabytes of main memory, which is sufficient to store all the data for many applications or at least to cache the applications' working sets [11] [38] [39].”
Kineograph: taking the pulse of a fast-changing and connected world (2012, background): “Distributed inmemory key/value stores have received a lot of attention, both in the research community and in the industry [13, 19, 21].”
Using RDMA efficiently for key-value services (2014, background): “They are used both as primary stores (e.g., Redis [4] and RAMCloud [23]), and as caches in front of backend databases (e.g., Memcached [5]).”