Safdar Jamil (ph.d.)

 Education

 Integrated MS - Ph.D. Student (March 2018 - February 2025) 

Dissertation: Deduplication Approaches for High-Performance and Space-Efficient Key-Value Databases

Advisor: Prof. Youngjae Kim

Sogang University, Seoul, South Korea

Bachelors of Computer Systems Engineering (January 2013 - December 2016)

Mehran University of Engineering and Technology, Pakistan (MUET)

Research Interests

Distributed Storage Systems, NoSQL Databases, and Distributed Deep Learning

Work Experience

• • • Post Doctoral Research Associate at Sogang University, Seoul, South Korea (March 2025 - Present)

      • Research Areas: Distributed Storage Systems, Near Storage Data Processing, and Distributed Deep Learning 

    • Graduate Research Assistant at Sogang University, Seoul, South Korea (March 2018 - February 2025)

      • Research Areas: NoSQL Databases, Storage Optimization, File Systems, and Data Structures and Algorithms

Research and Development Experience

Data Deduplication-enabled Log-Structure Merge (LSM) Tree-based Key-Value Store 

• Designing and developing a deduplication-enabled LSM-tree-based node-local distributed key-value, DENKV, which aimed to maintain the performance characteristics of LSM-tree while reducing the write and space amplification problems. 

• DENKV leverages the background thread pool and introduces the Asynchronous Partly Inline Deduplication (APID) module at the FLUSH operation and adopts a key-value separation design to maintain the structural constraint of the LSM-tree.

Tree-based Index Data Structures for Persistent-Memory-based Manycore Machines

• Designed and developed concurrent, non-uniform memory access (NUMA)-aware persistent B+-tree-based F3-tree and R-tree-based MPR-tree by exploiting the future objects and providing consistency guarantee through durable linearizability and lock-free read operations. 

Heap Memory Object Placement on Hybrid Memory System

• Modeled and developed an ILP-based heap memory object placement planner, eMAP, for a hybrid memory system comprised of DRAM and NVM. eMap takes into account the device characteristics such as read and writes latency and energy consumption. 

• In addition to initial memory object placement, an extended module of eMDyn monitors the energy consumption of the hybrid memory system and migrates the memory object from one memory device to another to meet the defined energy consumption threshold.

Skills and Expertise

Skillful in Linux, distributed systems, database systems, persistent memory-based data structures, and algorithms

• Programming Languages: C, C++, C#, Python 

• Non-Volatile Memory: Intel's PMDK Suite and Emulation platforms 

• Databases and Key-value Stores: RocksDB, SQLite

• Parallel and Distributed File Systems: Lustre

• Parallel Programming: MPI, pthread, std::future

• Benchmark: YCSB, FIO, IOR

• Tools: gcc, gdb, ctags, Git, Visual Studio, LaTex, Gnuplot, OmniGraffle

Seleted Publications

1. Safdar Jamil, Awais Khan, Xubin He, Youngjae Kim, DedupKV: A Space-Efficient and High-Performance Key-Value Store via Fine-Grained Deduplication, In Proceedings of the 39th ACM International Conference on Supercomputing, (ACM ICS) (2025), Salt Lake City, USA, June 8 - 11, 2025. [Paper]

2. Seoyeong Lee, Junghwan Park, Yoo-Chan Kim, Safdar Jamil, Awais Khan, Seung Woo Son, Jae-Kook Lee, So-Sik Ahn, Taeyoung Hong, Youngjae Kim, Leveraging Pre-Built Catalogs and Object-Level Scheduling to Eliminate I/O Bottlenecks, IEEE Access, March 2025. [Paper]

3. KiHwang Kim, Hyunsung Chung, Seonghoon Ahn, Junhyeok Park, Safdar Jamil, Hongsu Byun, Myungcheol Lee, Jinchun Choi, Youngjae Kim, KVAccel: A Novel Write Accelerator for LSM-Tree-Based KV Stores with Host-SSD Collaboration, In Proceedings of the 36th IEEE Int'l Parallel and Distributed Processing Symposium (IPDPS) (2025), Milan Italy, June 3-7, 2025. [Paper]

4. Safdar Jamil, Joseph Ro, Joo-Young Hwang, Youngjae Kim, Efficient Data Placement in Deduplication enable ZenFS via CRC-Based Prediction, IEEE Access, December 2024. [Paper]

5. Hamin Hwangbo, Joseph Ro, Sungjin Byeon, Safdar Jamil, Junyoung Han, Jooyoung Hwang, Youngjae Kim, Towards A Unified Garbage Collection Strategy in ZNS Key-Value Store File Systems Using Same-Victim GC, In Proceedings of the IEEE Int'l Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS) (2024), Krakow, Poland, October 21-23, 2024. [Paper] 

6. Hongsu Byun, Safdar Jamil, Jungwook Han, Sungyong Park, Myuncheol Lee, Changsoo Kim, Beongjun Choi, Youngjae Kim, An Analytical Model-based Capacity Planning Approach for Building CSD-based Storage Systems, ACM Transactions on Embedded Computing Systems (TECS), September 2023. [Paper]

7. Sungjin Byeon, Joseph Ro, Safdar Jamil, Jeon-Uk Kang, Youngjae Kim, A Free-Space Adaptive Runtime Zone-Reset Algorithm for Enhanced ZNS Efficiency, In the Proceedings of the 14th Workshop on Hot Topics in Storage and File Systems (HotStorage 2023), Boston, MA, USA, July 9, 2023. [Paper]

8. Jinhoon Lee, Yeonwoo Jung, Suyeon Lee, Safdar Jamil, Sungyong Park, Kwangwon Koh, Hongyeon Kim, Kangho Kim, Youngjae Kim, MFence: Defending Against Memory Access Interference in Disaggregated Cloud Memory Platform, In Proceedings of The 38th ACM/SIGAPP Symposium On Applied Computing (ACM SAC 2023), Tallinn Estonia, March 2023. [Paper]

9. Abdul Salam, Safdar Jamil, Sungwon Jung, Sung-Soon Park, Youngjae Kim, Future-based Persistent Spatial Data Structure for NVM-based Manycore Machines, IEEE Access, October 2022. [Paper]

10. Safdar Jamil, Abdul Salam, Awais Khan, Burgstaller Bernd, Sung-Soon Park, Youngjae Kim, Scalable NUMA-aware Persistent B+-Tree for Non-Volatile Memory Devices, Cluster Computing: The Journal of Networks, Software Tools and Applications, September 2022. [Paper]

11. Safdar Jamil, Awais Khan, Kihyun Kim, Jae-Kook Lee, Dosik Ahn, Taeyoung Hong, Sarp Oral, Youngjae Kim, DENKV: Addressing Design Trade-offs of Key-value Stores for Scientific Applications, In Proceedings of 7th International Parallel Data Systems Workshop (PDSW 2022) held in conjunction with SC22, Dallas, TX, November 2022. [Paper]

12. Safdar Jamil, Awais Khan, Youngjae Kim, Exploring Data Deduplication in LSM Tree-based Key-Value Stores,  Work-In-Progress, In Proceedings of USENIX Conference on File and Storage Technologies (FAST) (2022), San Jose, CA, February 2022. [Talk]

13. Safdar Jamil, Awais Khan, Burgstaller Bernd, Youngjae Kim, Towards Scalable Manycore-aware Persistent B+-Tree for Efficient Indexing in Cloud Environment,  In Proceedings of 9th International Workshop on Autonomic Management of High-performance Grid and Cloud Computing (AMGCC), Washington D.C., September 27, 2021. [Paper]

14. June-Hyung Kim, Youngjae Kim, Safdar Jamil, Chang-Gyu Lee, Sungyong Park, Parallelizing Shared File I/O operations of NVM File System for Manycore Servers, IEEE Access, 2021. [Paper]

15. June-Hyung Kim, Youngjae Kim, Safdar Jamil, Sungyong Park, A NUMA-aware NVM File System Design for Manycore Server Applications, (Short Paper) In Proceedings of the IEEE Int'l Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS) (2020), Nice, France, October 22-25, 2020. (Acceptance Rate: 27.0%) [Paper]

16. Joongeon Park, Safdar Jamil, Awais Khan, Matt Sangkeun Lee, Youngjae Kim, ScaleML: Machine Learning based Heap Memory Object Scaling Prediction, In Proceedings of the IEEE Non-Volatile Memory Systems and Applications (NVMSA) (2020), 19-21 August 2020. [Paper]

17. Taeuk Kim, Safdar Jamil, Joongeon Park, Youngjae Kim, Optimizing Heap Memory Object Placement in the Hybrid Memory System with Energy Constraints, IEEE Access, 2020. [Paper]

Awards and Honors

• USENIX FAST Student/Travel Grant, 2023

• ACSOS Student Grant, 2021

• Global Fellowship, Sogang University, 2018 - Present.

• Korea Software Congress Student grant, 2018.

• Microsoft Student Partner, 2015.

Ongoing works

1. Elastic Deduplication Controller for LSM-tree based Key-Value Stores (under submission)