• Corpus ID: 14797711

Tiered Replication: A Cost-effective Alternative to Full Cluster Geo-replication

@inproceedings{Cidon2015TieredRA,
  title={Tiered Replication: A Cost-effective Alternative to Full Cluster Geo-replication},
  author={Asaf Cidon and Robert Escriva and Sachin Katti and Mendel Rosenblum and Emin G{\"u}n Sirer},
  booktitle={USENIX Annual Technical Conference},
  year={2015}
}
Cloud storage systems typically use three-way random replication to guard against data loss within the cluster, and utilize cluster geo-replication to protect against correlated failures. This paper presents a much lower cost alternative to full cluster geo-replication. We demonstrate that in practical settings, using two replicas is sufficient for protecting against independent node failures, while using three random replicas is inadequate for protecting against correlated node failures. We… 
A popularity-aware cost-effective replication scheme for high data durability in cloud storage
TLDR
To handle both correlated and independent failures, PMCR stores the three replicas of the same data into one Copyset formed by two servers in the primary tier and one server in the backup tier, and uses the Delta Compression method for write-intensive data to reduce storage cost and bandwidth cost.
A Low-Cost Multi-failure Resilient Replication Scheme for High Data Availability in Cloud Storage
  • Jinwei Liu, Haiying Shen
  • Computer Science
    2016 IEEE 23rd International Conference on High Performance Computing (HiPC)
  • 2016
TLDR
A nonlinear integer programming model is built to maximize data availability in both types of failures and minimize the cost caused by replication, and shows that MRR achieves high data availability, low data loss probability and low consistency maintenance cost and storage cost compared to previous replication schemes.
Popularity-Aware Multi-Failure Resilient and Cost-Effective Replication for High Data Durability in Cloud Storage
TLDR
Extensive numerical results based on trace parameters and experimental results from real-world Amazon S3 show that PMCR achieves high data durability, low probability of data loss, and low storage cost and bandwidth cost compared to previous replication schemes.
RAIDP: replication with intra-disk parity
TLDR
A new point in the distributed storage design space between replication and erasure coding is proposed, RAIDP, which maintains only two replicas, rather than three or more, and is implemented in HDFS, which triplicates by default.
Proactive Re-replication Strategy in HDFS based Cloud Data Center
TLDR
This paper proposes a proactive re-replication strategy that uses predicted CPU utilization, predicted disk utilization, and popularity of the replicas to perform re-Replication effectively while ensuring all the server workloads are balanced.
Avoiding Performance Impacts by Re-Replication Workload Shifting in HDFS Based Cloud Storage
TLDR
This paper presents a deferred re-Replication algorithm to dynamically shift the re-replication workload based on current resource utilization status of the system, and can reduce performance impacts on users’ application workloads while ensuring the same reliability level as default HDFS can provide.
A DATA RE-REPLICATION SCHEME AND ITS IMPROVEMENT TOWARD PROACTIVE APPROACH
With increasing demand for cloud computing technology, cloud infrastructures are utilized to their maximum limits. There is a high possibility that commodity servers that are used in Hadoop
A control theoretical approach to non-intrusive geo-replication for cloud services
TLDR
The model provides a control-theoretical solution based on Model Predictive Control to dynamically allocate network bandwidth to accommodate the objectives of both replication and application data streams and is able to maintain a consistent performance over periods of persistent overload.
PRS: A Pattern-Directed Replication Scheme for Heterogeneous Object-Based Storage
TLDR
This article introduces a new data replication scheme called Pattern-directed Replication Scheme (PRS) to achieve efficient data replication for heterogeneous storage systems and evaluates the results with extensive tests in Sheepdog, a typical object-based storage system.
E2FS: an elastic storage system for cloud computing
TLDR
E2FS analyzes features of data and makes dynamic replication decision to balance the cost and performance of cloud storage, and can dynamically scale in/out the storage system based on real-time demands of big data applications.
...
1
2
3
4
...

References

SHOWING 1-10 OF 42 REFERENCES
Efficient Replica Maintenance for Distributed Storage Systems
TLDR
The paper proposes the Carbonite replication algorithm for keeping data durable at a low cost and shows that Carbonite is able to keep all data durable and uses 44% more network traffic than a hypothetical system that only responds to permanent failures.
Proactive Replication for Data Durability
TLDR
Tempo, a distributed hash table that allows each user to specify a maximum maintenance bandwidth and uses it to perform proactive replication, is introduced and results suggest that Tempo can deliver high durability despite only using several kilobytes per second of bandwidth, comparable to state-ofthe-art reactive systems.
Copysets: Reducing the Frequency of Data Loss in Cloud Storage
TLDR
Copyset Replication is presented, a novel general-purpose replication technique that significantly reduces the frequency of data loss events and presents a near optimal tradeoff between the number of nodes on which the data is scattered and the probability ofData loss.
Making Geo-Replicated Systems Fast as Possible, Consistent when Necessary
TLDR
This work proposes RedBlue consistency, which enables blue operations to be fast while the remaining red operations are strongly consistent (and slow), and introduces a method that increases the space of potential blue operations by breaking them into separate generator and shadow phases.
Availability in Globally Distributed Storage Systems
TLDR
This work characterize the availability properties of cloud storage systems based on an extensive one year study of Google's main storage infrastructure and presents statistical models that enable further insight into the impact of multiple design choices, such as data placement and replication strategies.
Stronger Semantics for Low-Latency Geo-Replicated Storage
TLDR
The evaluation shows that the Eiger system achieves low latency, has throughput competitive with eventually-consistent and non-transactional Cassandra, and scales out to large clusters almost linearly (averaging 96% increases up to 128 server clusters).
Chain Replication for Supporting High Throughput and Availability
TLDR
Besides outlining the chain replication protocols themselves, simulation experiments explore the performance characteristics of a prototype implementation and several object-placement strategies (including schemes based on distributed hash table routing) are discussed.
Glacier: highly durable, decentralized storage despite massive correlated failures
TLDR
Glasgow is described, a distributed storage system that relies on massive redundancy to mask the effect of large-scale correlated failures and is used as the storage layer for an experimental serverless email system.
Transaction chains: achieving serializability with low latency in geo-distributed storage systems
TLDR
It is shown that it is possible to obtain both serializable transactions and low latency, under two conditions: transactions are known ahead of time, permitting an a priori static analysis of conflicts, and transactions are structured as transaction chains consisting of a sequence of hops.
Kinesis: A new approach to replica placement in distributed storage systems
TLDR
Evaluations driven by real-world traces show that Kinesis can significantly outperform the widely used Chain replica-placement strategy in terms of resource requirements, end-to-end delay, and failure recovery.
...
1
2
3
4
5
...