Megastore: Providing Scalable, Highly Available Storage for Interactive Services

By J. Baker, C. Bond, J. Corbett, JJ Furman, Andrey Khorlin, James Larson, Jean-Michel Léon, Yawei Li, Alexander Lloyd, Vadim Yushprakh. Conference on Innovative Data Systems Research, 2011

830 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/3038a146f2ed61408865b5487173ead0da444a50

How researchers see this paper

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

Citations in the corpus

A Secure Sharding Protocol For Open Blockchains (2016, background): “Sharding protocols are commonly used in distributed databases and in cloud infrastructure, wherein certain network infrastructure can be trusted ( e.g. , see the commonly used two-phase commit protocol) or where the goal is to tolerate crash (non-byzantine) failures [15–17].”
Stronger Semantics for Low-Latency Geo-Replicated Storage (2013, unknown): “Traditional databases, as well as the more recent Walter [52], MDCC [35], Megastore [9], and some Cassandra configurations, provide stronger semantics and a rich data model, but cannot guarantee low latency.”
Making Smart Contracts Smarter (2016, background): “able to attempted manipulation by arbitrary adversaries— a threat that is restricted to accidental failures in traditional permissioned networks such as centralized cloud services [10, 11].”
Making Geo-Replicated Systems Fast as Possible, Consistent when Necessary (2012, unknown): “The rest of the paper is organized as follows: we po- yes yes N/A Timeline/snapshot PNUTS [8], Megastore [3] reads only yes yes yes yes N/A Fork SUNDR [24] all ops no yes yes yes N/A Eventual Bayou [38], Depot [26] all ops yes no yes yes N/A Sporc [12], CRDT [33] all ops yes yes no yes N/A Zeno [34]”
Calvin: fast distributed transactions for partitioned database systems (2012, methodology): “Google’s Megastore [7] and IBM’s Spinnaker [25] recently pi oneered the use of the Paxos algorithm [18, 19] for strongly co nsistent data replication in modern, high-volume transactiona l databases”
TAO: Facebook's Distributed Data Store for the Social Graph (2013, background): “Megastore is a storage system that uses Paxos across geographically distributed data centers to provide strong consistency guarantees and high availability [5].”
Don't settle for eventual: scalable causal consistency for wide-area storage with COPS (2011, background): “As noted earlier, CAP states that linearizable systems cannot have latency lower than their round-trip inter-datacenter latency; only recently have they been used for wide-area operation, and only when the low latency of ALPS can be sacrificed [9].”
There is more consensus in Egalitarian parliaments (2013, background): “Many databases are accessed simultaneously from different continents, requiring georeplication [2, 8].”
Performance Analysis of Cloud Computing Centers Using M/G/m/m+r Queuing Systems (2012, background): “Due to the dynamic nature of cloud environments, diversity of user's requests and time dependency of load, cloud centers must provide expected quality of service at widely varying loads [7], [8].”
Spanner: Google's Globally-Distributed Database (2012, unknown): “Many applications at Google have chosen to use Megastore [5] because of its semi-relational data model and support for synchronous replication, despite its relatively poor write throughput.”
Windows Azure Storage: a highly available cloud storage service with strong consistency (2011, background): “As in many other highly-available distributed storage systems [6,14,1,5], WAS also provides geo-redundancy.”
Tachyon: Reliable, Memory Speed Storage for Cluster Computing Frameworks (2014, background): “Practitioners and researchers have built a wide array of programming frameworks [34– 36, 42, 53, 54] and storage systems [14, 15, 25, 26, 39] tailored to a variety of workloads.”
Transactional storage for geo-replicated systems (2011, methodology): “Megastore [7] partitions data and provides the ACID properties within a partition but, unlike Walter, it fails to provide full transactional semantics for reads across partitions.”
Large-scale cluster management at Google with Borg (2015, methodology): “Our distributed storage systems such as GFS [34] and its successor CFS, Bigtable [19], and Megastore [8] all run on Borg.”
Toward Scalable Systems for Big Data Analytics: A Technology Tutorial (2014, background): “• Megastore [156] blends the scalability of NoSQL data stores with the convenience of traditional RDBMSs to achieve both strong consistency and high availability.”
Trends in big data analytics (2014, background): “Bigtable [21] andMegastore [13] are key-value stores that replicate”