Application of data fragmentation and replication methods in the cloud: a review

Castro-Medina, Felipe; Rodríguez-Mazahua, Lisbeth; Abud-Figueroa, María Antonieta; Romero-Torres, Celia; Reyes-Hernández, Luis Ángel; Alor-Hernández, Giner

doi:10.1109/conielecomp.2019.8673249

Cited by 11 publications

(3 citation statements)

References 69 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Globally, data replication in the environment is evolving as an explicit data management technique in the cloud environment [25]. In data replication environments, there are two (2) common mechanisms for replication strategies.…”

Section: Related Workmentioning

confidence: 99%

Replication Strategy with Comprehensive Data Center Selection Method in燙loud Environments

Fazlina¹,

Latip²,

Ibrahim³

et al. 2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

As the amount of data continues to grow rapidly, the variety of data produced by applications is becoming more affluent than ever. Cloud computing is the best technology evolving today to provide multi-services for the mass and variety of data. The cloud computing features are capable of processing, managing, and storing all sorts of data. Although data is stored in many high-end nodes, either in the same data centers or across many data centers in cloud, performance issues are still inevitable. The cloud replication strategy is one of best solutions to address risk of performance degradation in the cloud environment. The real challenge here is developing the right data replication strategy with minimal data movement that guarantees efficient network usage, low fault tolerance, and minimal replication frequency. The key problem discussed in this research is inefficient network usage discovered during selecting a suitable data center to store replica copies induced by inadequate data center selection criteria. Hence, to mitigate the issue, we proposed Replication Strategy with a comprehensive Data Center Selection Method (RS-DCSM), which can determine the appropriate data center to place replicas by considering three key factors: Popularity, space availability, and centrality. The proposed RS-DCSM was simulated using CloudSim and the results proved that data movement between data centers is significantly reduced by 14% reduction in overall replication frequency and 20% decrement in network usage, which outperformed the current replication strategy, known as Dynamic Popularity aware Replication Strategy (DPRS) algorithm.

show abstract

Section: Related Workmentioning

confidence: 99%

Replication Strategy with Comprehensive Data Center Selection Method in燙loud Environments

Fazlina¹,

Latip²,

Ibrahim³

et al. 2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

show abstract

“…A poor tail latency will impact most users in the case of large-scale applications with modern architectures, such as microservices, especially when a single user interaction can translate into many service calls. Improving the tail latency is a well-known performance engineering challenge that spans multiple layers of the cloud software stack, and it has been addressed by academics in different ways [3], [7], [16], [18], [23]. The majority of existing research efforts focus on high-level abstractions, neglecting the lower end of the cloud stack, such as the Operating System (OS) layer [10].…”

Section: Introduction and Related Workmentioning

confidence: 99%

Inducing Huge Tail Latency on a MongoDB deployment

Andreoli,

Cucinotta

2023

2023 IEEE International Conference on Cloud Engineering (IC2E)

View full text Add to dashboard Cite

The NoSQL paradigm has emerged as the leading design choice for cloud providers offering highly scalable storage services. Contrary to traditional relational databases, NoSQL architectures are capable of ingesting the ever-growing volume of nowadays' data-driven applications characterized by low-latency and high-throughput requirements. However, it is difficult to build an ultra-scalable, high-performance storage engine that can sustain an arbitrary number of concurrent clients. A common technique to increase throughput is minimizing the OS overhead, quantified as the number of context switches, through busy waiting (or "spinning"). While this simple synchronization mechanism proves to be beneficial in the high-performance computing community, it requires special care to avoid wasting resources and counter-intuitive behaviors. In this paper, we address an instance of "unsafe" busy waiting in WiredTiger, the underlying storage engine of MongoDB, which leads to a consistent, excessive increase of tail latency in high contention scenarios.

show abstract

“…Intelligent Fragmentation of the databases effectively establishes the cohesiveness between fragments and aligns the data to the users in perfect patterns. However, the traditional database fragmentation methods are insufficient as the latest distributed databases are large with numerous attributes and tuples [7]. The fragmentation methods are either horizontal or vertical fragmentation.…”

Section: Introductionmentioning

confidence: 99%

Hybrid Data Fragmentation Using Genetic Killer Whale Optimization-Based Clustering Model

2022

pnr

View full text Add to dashboard Cite

Effective storage structure is essential in distributed data processing systems as it determines the users' consistent, scalable and reliable data accessibility. A flexible distributed database system's major aspects are data fragmentation and subsequent allocation. An efficient data fragmentation technique must support both the horizontal and vertical categorization of the tuples so that the frequency and type of queries for the data statistics are learned along with the data pattern and the attributes that impact these patterns. However, most existing fragmentation methods are based only on the query frequency and type for the data statistics, leading to less effective analysis of the data patterns and attributes in large databases. To avoid the data loss and imperfect grouping of large databases, this paper presents a hybrid data fragmentation model using an advanced optimization-based clustering model for performing horizontal and vertical fragmentation. This proposed model, called Genetic Killer Whale Optimization based Clustering (GKWOC), is developed by integrating the Killer Whale Optimization (KWO) and Genetic Algorithms (GA) into the standard subspace clustering algorithm. This optimization-based hybrid fragmentation model performs fragmentation in terms of the tuples and the attributes. Since the databases can be high dimensional, subspace-based clustering can improvise good fragments for distributed databases and reduce the data access time for the users. This proposed model showed better fragmentation performance than the state-of-the-art methods during the evaluations, thus suitable for large distributed databases with competent access time and reduced error rates.

show abstract

Application of data fragmentation and replication methods in the cloud: a review

Cited by 11 publications

References 69 publications

Replication Strategy with Comprehensive Data Center Selection Method in燙loud Environments

Replication Strategy with Comprehensive Data Center Selection Method in燙loud Environments

Inducing Huge Tail Latency on a MongoDB deployment

Hybrid Data Fragmentation Using Genetic Killer Whale Optimization-Based Clustering Model

Contact Info

Product

Resources

About