ByteHTAP

Chen, Jianjun; Ding, Yonghua; Liu, Ye; Fang-shi, LI; Zhang, Li; Zhang, Mingyi; Wei, Kui; Cao, Lixun; Zhang, Dan; Liu, Yang; Zhang, Lei; Shi, Rui; Ding, Wei; Wu, Kai; Luo, Shangyu; Sun, Jason; Liang, Yu-Ming

doi:10.14778/3554821.3554832

Cited by 12 publications

(7 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…-[High Data Freshness] A significant motivation of HTAP is to eliminate ETL, thus providing intelligent insights into fresh data at generation speed. For that reason, freshness is commonly mentioned as an essential design goal of HTAP [39,58,78,83,103,111,150,154,168]. -[Strong Performance Isolation] Performance isolation refers to the ability to maintain the performance of a specified workload (e.g., OLTP) while another workload changes.…”

Section: Design Goals and Evaluation Metricsmentioning

confidence: 99%

“…consistency is always good to have. Multiple HTAP systems (e.g., [39,78,168]) target maintaining strong consistency across OLTP and OLAP. By doing so, application developers do not need to spend extra effort handling data consistency issues in the application layer, largely reducing the development complexity.…”

Section: -[Strong Consistency Across Oltp and Olap] Strongmentioning

confidence: 99%

“…This architecture represents the road to building an HTAP database on the shoulder of existing row and column stores. Typical implementations include TiDB [78], Heatweave [123], and F1 Lightning [175], IBM BLU [137], and ByteHTAP [39].…”

Section: A Taxonomy Of Existing Achitecturesmentioning

confidence: 99%

“…Then, they execute queries on the desirable engines and stores. Typical implementations are PolarDB-IMCI [168] and ByteHTAP [39]. Another approach handles the diversion problem with a unified optimizer (Fig.…”

Section: Query Optimization and Executionmentioning

confidence: 99%

See 3 more Smart Citations

A survey on hybrid transactional and analytical processing

Song,

Zhou,

Cui

et al. 2024

The VLDB Journal

View full text Add to dashboard Cite

To provide applications with the ability to analyze fresh data and eliminate the time-consuming ETL workflow, hybrid transactional and analytical (HTAP) systems have been developed to serve online transaction processing and online analytical processing workloads in a single system. In recent years, HTAP systems have attracted considerable interest from both academia and industry. Several new architectures and technologies have been proposed. This paper provides a comprehensive overview of these HTAP systems. We review recently published papers and technical reports in this field and broadly classify existing HTAP systems into two categories based on their data formats: monolithic and hybrid HTAP. We further classify hybrid HTAP into four sub-categories based on their storage architecture: row-oriented, column-oriented, separated, and hybrid. Based on such a taxonomy, we outline each stream’s design challenges and performance issues (e.g., the contradictory format demand for monolithic HTAP). We then discuss potential solutions and their trade-offs by reviewing noteworthy research findings. Finally, we summarize emerging HTAP applications, benchmarks, future trends, and open problems.

show abstract

Section: Design Goals and Evaluation Metricsmentioning

confidence: 99%

Section: -[Strong Consistency Across Oltp and Olap] Strongmentioning

confidence: 99%

Section: A Taxonomy Of Existing Achitecturesmentioning

confidence: 99%

Section: Query Optimization and Executionmentioning

confidence: 99%

See 2 more Smart Citations

A survey on hybrid transactional and analytical processing

Song,

Zhou,

Cui

et al. 2024

The VLDB Journal

View full text Add to dashboard Cite

show abstract

“…However, some real workloads require long transactions. The bill of materials workload requires long transactions [19], and transactional analytics for social data requires range queries of over 50,000 records [1].…”

Section: Introduction 1motivationmentioning

confidence: 99%

CLMD: Making Lock Manager Predictable and Concurrent for Deterministic Concurrency Control

Uchida,

Kawashima

2024

IJNC

View full text Add to dashboard Cite

Transaction processing is common in our daily lives, and various protocols have been studied to manage concurrency control. One such protocol is the deterministic concurrency control protocol, which is highly efficient in handling workloads with high contention because of its deterministic transaction scheduling nature. This paper introduces a new lock manager, the concurrent lock manager for determinism or CLMD. CLMD allows non-conflicting transactions to be executed concurrently on deterministic concurrency control protocols. The concept is to check for conflicts between the current and all future transactions while the conventional scheme checks only the current and the following ones. The CLMD eliminates the bottleneck present in the conventional locking scheme used in the original Calvin paper. We evaluated CLMD with Calvin in experiments. The results showed that the proposed method outperformed SS2PL under high-contention workloads and performed better than Calvin under low and highcontention workloads. The maximum performance improvement observed was approximately 44.4 times using 64 logical cores.

show abstract

EPO‐R: An efficient garbage collection scheme for long‐term transactions

Onishi,

Hoshino,

Kawashima

2024

Concurrency and Computation

View full text Add to dashboard Cite

SummaryThis article proposes EPO‐R, which is an efficient garbage collection scheme designed for multi‐version concurrency control (MVCC) protocols. MVCC generates a version for each update operation, and it can exhaust physical memory for dynamically changing environments such as IoT payments. Eager pruning of obsolete versions (EPO) is a novel garbage collection technique for long‐term transactions. We found room for improvement in EPO, which triggers reclamation invocation. EPO is triggered for each write operation, and it wastes CPU resources. The proposed method EPO‐R is triggered on a read operation to address this issue. The result of experiments with 64 CPU cores and workloads with the combination of short‐term and long‐term transactions demonstrated that EPO‐R exhibited 6.3 times higher throughput than that of EPO. The reason for this can be explained by analyzing the number of unreclaimable versions.

show abstract

ByteHTAP

Cited by 12 publications

References 33 publications

A survey on hybrid transactional and analytical processing

A survey on hybrid transactional and analytical processing

CLMD: Making Lock Manager Predictable and Concurrent for Deterministic Concurrency Control

EPO‐R: An efficient garbage collection scheme for long‐term transactions

Contact Info

Product

Resources

About