Detecting Isolation Anomalies in Relational DBMSs
Read the paper · doi:10.1145/3728953
What this paper does with SQLancer
SQLancer appears only in the future-work discussion, as background on what else DBMS testing covers: the paper notes that SQLancer and its approaches -- PQS, TLP and NoREC -- detect logic bugs in SELECT statements, that QPG uses query plans to guide testing, and that QuDi, GQT and GRev carry equivalent-query ideas to graph systems. That is a survey of the adjacent literature rather than a relationship: the paper's own comparison is against isolation checkers, and the anomaly work it builds on is Emme and TxCheck.
This work checks isolation levels in relational DBMSs directly, rather than through the key-value abstraction existing isolation checkers assume. Because those checkers model only read(key) and write(key,value), they cannot represent the SQL operations real workloads use: of 48 isolation anomalies the authors collect, existing checkers identify 2. The approach builds dependency graphs over relational operations and detects the cycles that indicate each anomaly class.
How it was classified
uses infrastructure — no
No reuse is described; the paper implements dependency-graph construction over relational operations.
extends technique — no
Isolation-anomaly detection through dependency-graph cycles shares no mechanism with the SELECT-oriented oracles the paper cites.
compares with — no
M7 states the comparison is with existing isolation checkers. SQLancer's oracles are mentioned only in the future-work survey, with no run reported.
describes as state of the art — no
The mentions describe what each technique does; no standing is claimed for any of them.
SQLancer publications it cites (8)
Bibliography entries that resolved to a SQLancer publication, or to a paper by one of the project's authors. A sentence citing one of these numbers is a reference to SQLancer even when it never writes the name.
| # | Entry | Matched as |
|---|---|---|
| 12 | 2024. SQLancer. https://www.manuelrigger.at/dbms-bugs. | sqlancer publication |
| 18 | Jinsheng Ba and Manuel Rigger. 2023. Testing Database Engines via Query Plan Guidance. In Proceedings of IEEE/ACM International Conference on Software Engineering (ICSE). 2060–2071. | sqlancer publication · QPG |
| 30 | Jack Clark, Alastair F Donaldson, John Wickerson, and Manuel Rigger. 2024. Validating Database System Isolation Level Implementations with Version Certificate Recovery. In Proceedings of European Conference on Compute... | project authored |
| 43 | Yuancheng Jiang, Jiahao Liu, Jinsheng Ba, Roland H. C. Yap, Zhenkai Liang, and Manuel Rigger. 2024. Detecting Logic Bugs in Graph Database Management Systems via Injective and Surjective Graph Query Transformation. In... | project authored |
| 44 | Zu-Ming Jiang, Si Liu, Manuel Rigger, and Zhendong Su. 2023. Detecting Transactional Bugs in Database Engines via Graph-Based Oracle Construction. In Proceedings of USENIX Symposium on Operating Systems Design and Imp... | project authored |
| 51 | Manuel Rigger and Zhendong Su. 2020. Detecting Optimization Bugs in Database Engines via Non-Optimizing Reference Engine Construction. In Proceedings of ACM Joint European Software Engineering Conference and Symposium... | sqlancer publication · NOREC |
| 52 | Manuel Rigger and Zhendong Su. 2020. Finding Bugs in Database Systems via Query Partitioning. Proceedings of the ACM on Programming Languages 4, OOPSLA (2020), 211:1–211:30. | sqlancer publication · TLP |
| 53 | Manuel Rigger and Zhendong Su. 2020. Testing Database Engines via Pivoted Query Synthesis. In Proceedings of USENIX Symposium on Operating Systems Design and Implementation (OSDI). 667–682. | sqlancer publication · PQS |
Every place it refers to SQLancer (14)
14 sentences, each stored verbatim from the extracted text with where it was found and how. “Citation marker” means the sentence names no tool at all and was reached through a reference number that resolved to a SQLancer publication.
| Id | Sentence | Found by | Where |
|---|---|---|---|
| M1 | Existing isolation checkers [ 24,27,28,30,42,46,49,58,62] mainly work on 𝑘𝑒𝑦-𝑣𝑎𝑙𝑢𝑒 -like data models, in which data is organized in a 𝑘𝑒𝑦-𝑣𝑎𝑙𝑢𝑒 format. |
citation marker project authored |
1 Introduction page 2 |
| M2 | Existing isolation checkers are designed specifically for 𝑘𝑒𝑦-𝑣𝑎𝑙𝑢𝑒 -like data models [ 24,27,28,30,42,46,49,58,62], and are not equipped to handle the complexities of relational data models in relational DBMSs. |
citation marker project authored |
1 Introduction page 3 |
| M3 | Existing isolation checkers [ 24,27,28,30,42,46,49,58,62] are limited to supporting simple read and write operations based on 𝑘𝑒𝑦, i. |
citation marker project authored |
1 Introduction page 3 |
| M4 | For these 48 isolation anomalies, existing isolation checkers [ 24,27,28,30,42,46,49,58,62] can only identify 2 anomalies due to their incapability of handling relational data models and complex SQL operations. |
citation marker project authored |
1 Introduction page 3 |
| M5 | However, existing isolation checkers [ 24,27,28,30,42,46,49,58,62] are limited to 𝑘𝑒𝑦-𝑣𝑎𝑙𝑢𝑒 -like data models and the associated 𝑟𝑒𝑎𝑑(𝑘𝑒𝑦)and𝑤𝑟𝑖𝑡𝑒(𝑘𝑒𝑦,𝑣𝑎𝑙𝑢𝑒) operations. |
citation marker project authored |
2.1 Motivating Example page 4 |
| M6 | Existing works [ 16,17,30,42, 46,49,58,62] have proved that the appearance of different kinds of cycles in the dependency graph 𝐺indicates different kinds of isolation anomalies, e. |
citation marker project authored |
3.4 Isolation Anomaly Detection page 15 |
| M7 | 4 Comparison with Existing Isolation Checkers Existing isolation checkers [ 24,27,28,30,42,46,49,58,62] mainly work on 𝑘𝑒𝑦-𝑣𝑎𝑙𝑢𝑒 -like data models, in which data is organized in a 𝑘𝑒𝑦-𝑣𝑎𝑙𝑢𝑒 format. |
citation marker project authored |
5.4 Comparison with Existing Isolation Checkers page 18 |
| M8 | Existing isolation anomaly detection algorithms [ 24,27,28,30,42,46,49,58,62] also aim to detect cycles in the dependency graph, among which there are some efficient anomaly detection algorithms, e. |
citation marker project authored |
6.2 Future Work page 19 |
| M9 | Emme [30] is a white-box checker that retrieves transaction timestamps and recovers a version certificate to check isolation levels in DBMSs that use multi-version concurrency control timestamp ordering protocol to guarantee serializability. |
citation marker project authored |
6.2 Future Work page 20 |
| M10 | TxCheck [44] constructs semantically equivalent test cases and compares their execution results to detect transaction bugs. |
citation marker project authored |
6.2 Future Work page 20 |
| M11 | propose SQLancer [ 12] and several approaches, e. |
name |
6.2 Future Work page 20 |
| M12 | , PQS [ 53], TLP [ 52] and NoREC [ 51], to detect logic bugs in SELECT statements. |
technique |
6.2 Future Work page 20 |
| M13 | QPG [ 18] utilizes query plans to guide DBMS testing. |
technique |
6.2 Future Work page 20 |
| M14 | Qudi [ 63], GQT [ 43] and GRev [ 50] further utilize equivalent queries [ 59] to detect logic bugs in graph DBMSs. |
citation marker project authored |
6.2 Future Work page 20 |