[GBASE Event] ICDIS Database Parameter Tuning Competition Opens for Registration
Registration for the ICDIS Database Parameter Tuning Competition, co-hosted by General Data Technology Co., Ltd., the Guangdong Provincial Key Laboratory of Novel Security Intelligence Technologies, and Harbin Institute of Technology (Shenzhen), begins on July 14. The competition will focus on the open-source database MySQL and parameter tuning technologies based on the GBase database platform. The event is part of ICDIS 2022 (The 4th International Conference on Data Intelligence and Security), which will be held in Shenzhen, China, from August 24 to 26, 2022.
Background
Traditional database technologies often rely on heuristic algorithms or manual intervention for tasks such as parameter tuning, fault diagnosis, and index recommendation. However, in the big data era, the sheer number of database instances, increasingly complex scenarios, and massive data volumes have made traditional approaches inadequate. For example, cloud databases can have millions of instances, each with widely varying application scenarios and user proficiency levels. Directly applying traditional heuristic algorithms yields unsatisfactory results, and manual intervention cannot manage this scale. Machine learning techniques, with their powerful learning capabilities, are showing growing potential and application prospects in the database field.
Database configurations involve hundreds of adjustable system parameters that govern various aspects of database components. Configuration tuning techniques aim to enhance database adaptability to a given scenario by selecting an optimal parameter combination—for instance, allocating appropriate memory space to different caches to minimize disk I/O, or updating optimizer estimation metrics to improve plan selection. With algorithmic advances, increased computing power, and big data transformation, deep learning and reinforcement learning are opening up even greater optimization opportunities for database technology.
Schedule
Registration and Certification (July 14–July 31, UTC+8):
1. Registration: Visit the official competition website, complete your personal information registration to sign up.
2. Participants can form teams of 1 to 5 people; each participant may only join one team.
3. Participants must ensure registration information is accurate and valid. The organizing committee reserves the right to disqualify teams that do not meet the requirements and revoke any awards.
4. The deadline for registration and team change operations is 8:00 AM on July 31.
5. Participant and team information must be submitted via Tencent Docs.
6. Scan the QR code to join the official competition QQ group for the latest announcements.
Preliminary Round: 2022/07/16 8:00 – 2022/07/31 8:00 (UTC+8)
Details: Please refer to the "Problem Description" section for preliminary round details. Participants can submit results during the preliminary round, and evaluations will be conducted by the assessment system. The judging panel will review submitted results, select the top 20 teams to advance to the final, and announce the list.
One code submission opportunity per day, with execution logs provided.
Final Round: 2022/08/04 8:00 – 2022/08/18 23:59 (UTC+8)
Details: Please refer to the "Problem Description" section for final round details. Participants can submit results during the final round, and evaluations will be conducted by the assessment system. The judging panel will review submitted results and select the 6 winning teams, which will then be announced.
Two code submission opportunities per day, with execution logs provided.
Final Presentation: 2022/08/24 8:00 – 2022/08/26 18:00 (UTC+8)
1. Material Preparation: Teams advancing to the final presentation must prepare evaluation materials in advance, including code, solution description, and PPT slides.
2. Review Session: Will be conducted as a presentation session; detailed arrangements will be announced separately.
3. Final Ranking: The organizing committee will determine the final awards based on code, historical scores, and judge evaluations, and will present prizes and certificates.
Eligibility
Open to the public, including individuals from universities, research institutions, internet companies, etc.
Awards
First Prize: 1 team, RMB 10,000 per team, plus an award certificate.
Second Prize: 2 teams, RMB 5,000 per team, plus an award certificate.
Third Prize: 3 teams, RMB 2,000 per team, plus an award certificate.
Excellence Award: Top 20 teams in the preliminary round, with an award certificate.
Participation Award: Every participant will receive a customized Kingston 32GB USB flash drive.
Referral Program
From the competition launch, participants who successfully invite friends to register will receive referral rewards. Gifts are cumulative and unlimited as long as the invitation criteria are met.
Referral Reward — Custom competition laptop backpack Requirement — Invite 4 teams to register
Organization
Hosts:
Guangdong Provincial Key Laboratory of Novel Security Intelligence Technologies
General Data Technology Co., Ltd.
Harbin Institute of Technology (Shenzhen)
Co-organizers:
Guangdong Computer Society
Shenzhen Computer Society
HIT (Shenzhen) – General Data Technology Joint Laboratory on Database Technology
Steering Committee:
Huang Xuan, Secretary-General, Guangdong Computer Society
Wang Xuan, Professor, Harbin Institute of Technology (Shenzhen); Director, Guangdong Provincial Key Laboratory of Novel Security Intelligence Technologies
Mao Rui, Professor, Shenzhen University
Tang Yong, Professor, South China Normal University
Ding Mingfeng, Chairman, General Data Technology Co., Ltd.
Organizing Committee:
Luo Wenjian, Professor, Harbin Institute of Technology (Shenzhen)
Qi Shuhan, Assistant Professor, Harbin Institute of Technology (Shenzhen); Deputy Director, HIT (Shenzhen) – General Data Technology Joint Laboratory on Database Technology
Liu Yang, Assistant Professor, Harbin Institute of Technology (Shenzhen); Deputy Director, HIT (Shenzhen) – General Data Technology Joint Laboratory on Database Technology
Yang Weiwei, Director of the Chief Engineer Office, General Data Technology Co., Ltd.; Deputy Director, HIT (Shenzhen) – General Data Technology Joint Laboratory on Database Technology
Expert Committee:
Xia Wen, Associate Professor, Harbin Institute of Technology (Shenzhen)
Zhao Wei, CTO, General Data Technology Co., Ltd.
Wang Xuesong, Director of R&D, General Data Technology Co., Ltd.
Cui Zhiwei, Director of Product, General Data Technology Co., Ltd.
Zhang Jiajia, Associate Researcher, Harbin Institute of Technology (Shenzhen)
Tang Bo, Assistant Professor, Southern University of Science and Technology
Qin Jianbin, Distinguished Professor, Shenzhen University
Rules and Notices
Mobile submission and score viewing are not supported. Please use a PC for submission and viewing.
All forms of cheating and misconduct are strictly prohibited, including but not limited to plagiarism and malicious attacks. Once verified, the competition organizer may restrict participation or directly remove rankings. Please refer to the "Cheating and Misconduct Policy" for details.
The competition organizer reserves the right to adjust or modify competition rules, prize information, timelines, and all related content according to operational needs and other requirements. The final confirmation from the organizer shall prevail.
Preliminary Round Problem Description:
Participants should use parameter tuning methods to improve the throughput (TPS) of DBMS (MySQL 5.6) under a specified workload. Methods may include, but are not limited to, experience-based script tuning, heuristic tuning, traditional machine learning–based tuning, and reinforcement learning–based tuning. To help participants quickly become familiar with database tuning and its application, the organizer will provide a dedicated database tuning system image. (See the problem description for details.)
Final Round Problem Description:
This final round uses GBase 8s, General Data Technology's enterprise-grade distributed transactional database, as the database to be tuned. Participants should apply a tuning method to improve GBase 8s throughput (TPS) under a specified workload. Methods may include, but are not limited to, experience-based script tuning, heuristic tuning, traditional machine learning–based tuning, and reinforcement learning–based tuning. To help participants quickly become familiar with database tuning and its application, the organizer will provide a dedicated database tuning system image. (See the problem description for details.)
Cheating and Misconduct Policy:
Cheating and misconduct violate the principles of fair competition by improperly gaining a competitive advantage. Categories include:
(1) Plagiarism: Any act that fails to ensure originality is considered plagiarism. For example:
● Using others' results without citing the source in the reference.
● Continuously quoting others' code exceeding 30% of total code volume.
● Directly using ideas, algorithms, parameter configurations, or other content from blogs, GitHub, or other open-source platforms and channels without any original modification or adaptation.
● Other content besides citations is insufficient to support the main contribution and value of the solution or algorithm.
● Other situations where others' achievements are used directly as value output rather than as intermediate tools or materials for the solution (e.g., known conditions, data cleaning methods).
● Colluding with other teams or non-team members to copy each other's work.
(2) Unfair Competition: Acts that maliciously pursue high scores contrary to the competition's purpose are considered unfair competition. For example:
● Participants who can access competition-related data through channels other than the designated competition method.
● Inviting non-team members to participate in solving problems or designing solutions, or completing tasks through outsourcing, help-seeking, etc., outside the participating team.
● Using any data other than the competition-provided data in any form (unless explicitly permitted on the competition homepage).
● Manually labeling part or all of the test set and submitting it as a result.
● Using computing resources beyond those specified in the competition (unless no rules are specified on the homepage).
● Using multiple accounts to participate in the same competition.
● Exploiting platform or rule loopholes.
● Inviting non-team members to solve problems or design solutions, or completing tasks through outsourcing, help-seeking, etc., outside the participating team.
(3) Malicious Destruction: Any act of deliberately damaging the competition platform, evaluation system, or environment. For example:
● Deliberately uploading files containing viruses.
● Deliberately launching attacks against the competition platform or evaluation system to disrupt the competition order.
Handling Process:
Appeal
1. Teams deemed to have cheated have the right to appeal; email: dbtune@cs.hitsz.edu.cn
2. When appealing, the following must be submitted (including but not limited to):
● Competition screenshots
● Detailed explanation of the reason
● Calculation code for submitted results
3. Appeals must be submitted within 24 hours of being determined as cheating; late submissions will not be accepted.
4. The appeal result will be given within 24 hours, but the reason for the decision will not be disclosed. Each team is allowed only one appeal. If the appeal is denied, the cheating determination stands and will not be reconsidered.
5. Teams whose appeals are approved will have their participation rights immediately restored and their ranking reinstated.
6. The competition's expert judging panel holds the final authority to interpret appeal results.