- File systems and in-storage computing (hardware/software co-design).
+ File systems and in-storage computing.
@@ -189,84 +195,112 @@
Biography
Publications
- A Spatio-Temporal Series Data Model with Efficient Indexing and Layout for Cloud-Based Trajectory Data Management.
- Yang Guo, Zhiqi Wang, Jin Xue, and Zili Shao.
- The 40th International Conference on Data Engineering (ICDE 2024)(CCF-A).
- [code] [paper]
-
-
- MirrorKV: An Efficient Key-Value Store on Hybrid Cloud Storage with Balanced Performance of Compaction and Querying.
+ MirrorKV: An Efficient Key-Value Store on Hybrid Cloud Storage with Balanced Performance of Compaction and Querying. Zhiqi Wang, and Zili Shao. The 43rd ACM SIGMOD International Conference on Management of Data (SIGMOD 2024)(CCF-A).
[paper]
- Lightning Talk: Model, Framework and Integration for In-Storage Computing with Computational SSDs.
- Tianyu Wang, Jin Xue, Zelin Du, Zhiqi Wang, Yaotian Cui, and Zili Shao.
- The 60th ACM/IEEE Design Automation Conference (DAC 2023)(CCF-A)(invited paper).
- [paper]
-
-
- ForestTI: A Scalable Inverted-Index-Oriented Timeseries Management System with Flexible Memory Efficiency.
+ ForestTI: A Scalable Inverted-Index-Oriented Timeseries Management System with Flexible Memory Efficiency. Zhiqi Wang, and Zili Shao. The 42nd ACM SIGMOD International Conference on Management of Data (SIGMOD 2023)(CCF-A).
[code] [paper]
- BSCache: A Brisk Semantic Caching Scheme for Cloud-based Performance Monitoring Timeseries Systems.
- Kai Zhang, Zhiqi Wang, and Zili Shao.
- Proceedings of the 51st International Conference on Parallel Processing (ICPP 2022)(CCF-B).
- [code] [paper]
-
-
- TimeUnion: An Efficient Architecture with Unified Data Model for Timeseries Management Systems on Hybrid Cloud Storage.
+ TimeUnion: An Efficient Architecture with Unified Data Model for Timeseries Management Systems on Hybrid Cloud Storage. Zhiqi Wang, and Zili Shao. The 41st ACM SIGMOD International Conference on Management of Data (SIGMOD 2022)(CCF-A).
[code] [paper]
- TagTree: Global Tagging Index with Efficient Querying for Time Series Databases.
- Jin Xue, Zhiqi Wang, and Zili Shao.
- The 36th IEEE International Parallel & Distributed Processing Symposium (IPDPS 2022)(CCF-B).
- [code] [paper]
-
-
- Heracles: An Efficient Storage Model and Data Flushing for Performance Monitoring Timeseries.
+ Heracles: An Efficient Storage Model and Data Flushing for Performance Monitoring Timeseries. Zhiqi Wang, Jin Xue, and Zili Shao. The 47th International Conference on Very Large Data Bases (VLDB 2021)(CCF-A), Volume 14(6), 1080-1092.
[code] [paper]
-
-
-
Experience
-
- 11/2023 - Present: Postdoc in CUHK. In-storage computation research.
+ A Spatio-Temporal Series Data Model with Efficient Indexing and Layout for Cloud-Based Trajectory Data Management.
+ Yang Guo, Zhiqi Wang, Jin Xue, and Zili Shao.
+ The 40th International Conference on Data Engineering (ICDE 2024)(CCF-A).
+ [code] [paper]
+
+
+ Lightning Talk: Model, Framework and Integration for In-Storage Computing with Computational SSDs.
+ Tianyu Wang, Jin Xue, Zelin Du, Zhiqi Wang, Yaotian Cui, and Zili Shao.
+ The 60th ACM/IEEE Design Automation Conference (DAC 2023)(CCF-A)(invited paper).
+ [paper]
- 06/2022 - 08/2022: Huawei Cloud Database Innovation Lab internship. Optimization of the storage engine of OpenGemini.
+ BSCache: A Brisk Semantic Caching Scheme for Cloud-based Performance Monitoring Timeseries Systems.
+ Kai Zhang, Zhiqi Wang, and Zili Shao.
+ Proceedings of the 51st International Conference on Parallel Processing (ICPP 2022)(CCF-B).
+ [code] [paper]
- 06/2019 - 08/2019: Google Summer of Code 2019. Optimization of the storage engine of Prometheus.
+ TagTree: Global Tagging Index with Efficient Querying for Time Series Databases.
+ Jin Xue, Zhiqi Wang, and Zili Shao.
+ The 36th IEEE International Parallel & Distributed Processing Symposium (IPDPS 2022)(CCF-B).
+ [code] [paper]
-
Projects
+
Working Experience
- Gemini: A monolithic software/hardware co-design key-value file system with computational storage. First, it contains a host-side kernel file system, which translates the file semantics to key-value commands. Second, we customize the Linux NVMe driver to bypass the Linux block layer and transmit the key-value commands. Third, we carefully design the flash translation layer (FTL) in our real hardware platform (a computational SSD) to handle the received key-value commands and manage the physical area of the SSD.
+
+
Postdoctoral Fellow in CUHK
+
11/2023 - Present
+
+
+
+
+
Huawei Cloud Database Innovation Lab (internship)
+
06/2022 - 08/2022
+
- MirrorKV: An LSM-tree-based key-value store tailored for cloud storage (EBS, S3). The key idea is to design different compaction mechanisms for different storage tiers, and manage keys and values in two mirrored LSM-trees to maintain the data locality. This project derives from RocksDB (C++).
+
+
Google Summer of Code (internship)
+
06/2019 - 08/2019
+
+
+
+
Research Experience
+Big Data Systems
+Supported by Hong Kong General Research Fund: Optimizing Storage System Design for Spatial-Temporal Big Data (RGC Ref No. 15224918). Serve as the project participant.
+
- ForestTI: A memory-efficient timeseries storage engine. The key idea is to design a flexible inverted index that can dynamically alter the structure based on the memory pressure. This project derives from TimeUnion.
+ Timeseries Management Systems
+ A thorough research on the main design decisions of the timeseries management systems, including the data model, memory data management, and persistent data management.
+
+
Data model: To solve the data redundancy issue of the timeseries data from the same data source, we propose a unified data model for both tags and data samples of timeseries, with a novel compression mechanism and a two-level indexing design.
+
Memory data management: To mitigate the memory overhead and maintain more timeseries with limited memory, we design a flexible inverted index that can dynamically adapt its structure to the memory pressure.
+
Persistent data management: To achieve high insertion throughput of big timeseries data, we design a dynamic time-partitioned LSM-tree with high insertion throughput, decent space efficiency, and efficient out-of-order data handling.
+
+
+
+Storage Engines
+Supported by Hong Kong General Research Fund: StoreLess: Eliminating Redundancy for LSM-tree based Key-Value Stores as Database Storage Engines in Internet Applications (RGC Ref No. 14219422). Serve as the project participant.
+
- TimeUnion: A timeseries storage engine tailored for cloud storage (EBS, S3). First, it proposes a unified data model for timeseries tag management. Second, it presents a time-partitioned LSM-tree with hot/cold data separation and efficient out-of-order data handling. This project is written from scratch in C++.
+ LSM-Tree-Based Key-Value Stores with Hybrid Cloud Storage
+ LSM-tree-based key-value stores are widely used as the storage engines of big data systems. As the data volume scales up, it is a natural trend to deploy the system on the cloud. However, the existing LSM-tree designs can not adapt to cloud storage because of the huge performance gap. We design MirrorKV with a balanced read/write performance which separates keys and values into two mirrored LSM-trees for better data locality and read performance, and designs different compaction mechanisms for fast and slow storage to improve write performance.
+
+
+File Systems and In-Storage Computing
+Supported by Hong Kong General Research Fund Project: Data Model and Programming Framework for Function Offloading in In-SSD Computing (RGC Ref No. 14202123). Serve as the project participant.
+
- Heracles: A timeseries storage engine with group data management and efficient data flushing mechanism. This project derives from the storage engine of Prometheus (Golang).
+ A Monolithic Software/Hardware Co-Design Key-Value File System
+ To mitigate the metadata manipulation overhead and I/O amplification of the traditional file systems designed for block storage, we implement a file system with a key-value interface, which offloads the data management to our computational storage platform.
+
+
Host-side key-value filesystem: It translates the file semantics (inode and page contents) to key-value commands correspondingly.
+
Host storage communication: We customize the Linux NVMe driver to bypass the Linux block layer and transmit the key-value commands.
+
Storage-side design: We carefully design the flash translation layer (FTL) to handle the received key-value commands and manage the physical area of the SSD.
+
@@ -280,9 +314,9 @@
Awards
-
Teaching
+
Teaching Experience
+CSCI3150: Introduction to Operating Systems
- CSCI3150: Introduction to Operating Systems
Fall 2019
@@ -294,7 +328,7 @@
Teaching
-
Professional Activities
+
Professional Experience
Participation & Talks
@@ -308,11 +342,19 @@
Professional Activities
External Reviewer
-
2024: CODES+ISSS
-
2023: TODS
-
2022: DAC, CODES+ISSS, SIGPLAN/SIGBED
-
2021: DAC, ICCAD, CODES+ISSS
-
2020: DAC, ICCAD, CODES+ISSS
+
Journal
+
+
ACM Transactions on Database Systems (TODS)
+
IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems (TCAD)
+
+
Conference
+
+
Design Automation Conference (DAC)
+
International Conference on Computer Design (ICCD)
+
Design Automation and Test in Europe Conference (DATE)
+
International Conference on Computer Aided Design (ICCAD)
+
Conference on Languages, Compilers, and Tools for Embedded Systems (LCTES)