Safdar Jamil
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Safdar Jamil has a total of 41 co-authors. Here are the top co-authors by number of shared publications:
- Youngjae Kim
Sogang University
11 shared publications - Awais Khan
Oak Ridge National Laboratory, Oak Ridge, TN, USA
4 shared publications - Sungyong Park
Department of Computer Science and Engineering, Sogang University, Seoul, Republic of Korea
4 shared publications - Abdul Salam
Department of Computer Science and Engineering, Sogang University, Seoul, Republic of Korea
2 shared publications - Bernd Burgstaller
Yonsei University, Seoul, Republic of Korea
2 shared publications - Chang Soo Kim
Electronics and Telecommunications Research Institute, Daejeon, South Korea
2 shared publications
Latest Publications
- Open Access
- 8 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 580-595
DEDUPKV: A Space-Efficient and High-Performance Key-Value Store via Fine-Grained Deduplication
- 8 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 580-595
Open AccessDEDUPKV: A Space-Efficient and High-Performance Key-Value Store via Fine-Grained Deduplication
- Open Access
- 20 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 55984-55995
Leveraging Pre-Built Catalogs and Object-Level Scheduling to Eliminate I/O Bottlenecks in HPC Environments
Modern High-Performance Computing (HPC) environments face mounting challenges due to the shift from large to small file datasets, combined with an increasing number of users and parallelized applications. As HPC systems rely on Parallel File Systems (PFS), such as Lustre for data processing, performance bottlenecks stemming from Object Storage Target (OST) contention have become a significant concern. Existing solutions, such as LADS with its object-level scheduling approach, fall short in large-scale HPC environments due to their inability to effectively address metadata I/O bottlenecks and the growing number of I/O processes. This study highlights the pressing need for a comprehensive solution that tackles both OST contention and metadata I/O challenges in diverse HPC workloads. To address these challenges, we propose SwiftLoad, an object-level I/O scheduling framework that leverages a metadata catalog to improve the performance and efficiency of parallel HPC utilities. The adoption of the metadata catalog mitigates the metadata I/O bottlenecks that commonly occur in HPC utilities, a challenge that is particularly pronounced in object-level I/O scheduling. SwiftLoad addresses OST contention and the uneven distribution of I/O processes across different OSTs through mathematical modeling and incorporates a Loader Configuration Module to regulate the number of I/O processes. Evaluated with two representative utilities—data deduplication profiling and data augmentation—SwiftLoad demonstrated performance improvements of up to 5.63× and 11.0×, respectively, on a production supercomputer.
- 20 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 55984-55995
Open AccessLeveraging Pre-Built Catalogs and Object-Level Scheduling to Eliminate I/O Bottlenecks in HPC Environments
Modern High-Performance Computing (HPC) environments face mounting challenges due to the shift from large to small file datasets, combined with an increasing number of users and parallelized applications. As HPC systems rely on Parallel File Systems (PFS), such as Lustre for data processing, performance bottlenecks stemming from Object Storage Target (OST) contention have become a significant concern. Existing solutions, such as LADS with its object-level scheduling approach, fall short in large-scale HPC environments due to their inability to effectively address metadata I/O bottlenecks and the growing number of I/O processes. This study highlights the pressing need for a comprehensive solution that tackles both OST contention and metadata I/O challenges in diverse HPC workloads. To address these challenges, we propose SwiftLoad, an object-level I/O scheduling framework that leverages a metadata catalog to improve the performance and efficiency of parallel HPC utilities. The adoption of the metadata catalog mitigates the metadata I/O bottlenecks that commonly occur in HPC utilities, a challenge that is particularly pronounced in object-level I/O scheduling. SwiftLoad addresses OST contention and the uneven distribution of I/O processes across different OSTs through mathematical modeling and incorporates a Loader Configuration Module to regulate the number of I/O processes. Evaluated with two representative utilities—data deduplication profiling and data augmentation—SwiftLoad demonstrated performance improvements of up to 5.63× and 11.0×, respectively, on a production supercomputer.
- Open Access
- 19 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 197233-197246
Efficient Data Placement in Deduplication Enabled ZenFS via CRC-Based Prediction
The Zoned Namespace (ZNS) interface shifts data management responsibility to upper-level applications, requiring them to reclaim space by issuing the zone-reset command to ZNS SSD devices, a process known as garbage collection (GC). Application-level GC can lead to performance degradation due to the high valid data copy overhead, which is further exacerbated by the larger GC units in ZNS SSDs. However, the impact of larger GC units can be mitigated if GC operations are made interruptible, allowing I/O requests to be served during zone resets or block reclamation. Moreover, the adoption of offline data deduplication as a storage optimization technique in ZNS-based file systems like ZenFS presents additional challenges. Offline deduplication must consider lifetime-based file allocation to avoid deduplicating hot data, and placing unique and duplicate data blocks together can further increase valid data copy overhead during GC. To address these issues, we propose DeZNS , an innovative data placement strategy for deduplication-enabled ZenFS. DeZNS tackles the increased valid data copy overhead during GC in offline deduplication by employing a lightweight CRC32 checksum-based method to predict potential duplicates with minimal performance impact, segregating unique and duplicate data blocks. This segregation reduces valid data migration overhead during GC, while the interruptible GC mechanism ensures that ongoing I/O requests are not delayed during zone resets, maintaining ZenFS performance. Additionally, DeZNS integrates an offline deduplication module that operates on segregated zones. Our extensive evaluation shows that DeZNS reduces valid data migration by 28% compared to baseline ZenFS and by up to 2× compared to naive offline deduplication in microbenchmarks.
- 19 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 197233-197246
Open AccessEfficient Data Placement in Deduplication Enabled ZenFS via CRC-Based Prediction
The Zoned Namespace (ZNS) interface shifts data management responsibility to upper-level applications, requiring them to reclaim space by issuing the zone-reset command to ZNS SSD devices, a process known as garbage collection (GC). Application-level GC can lead to performance degradation due to the high valid data copy overhead, which is further exacerbated by the larger GC units in ZNS SSDs. However, the impact of larger GC units can be mitigated if GC operations are made interruptible, allowing I/O requests to be served during zone resets or block reclamation. Moreover, the adoption of offline data deduplication as a storage optimization technique in ZNS-based file systems like ZenFS presents additional challenges. Offline deduplication must consider lifetime-based file allocation to avoid deduplicating hot data, and placing unique and duplicate data blocks together can further increase valid data copy overhead during GC. To address these issues, we propose DeZNS , an innovative data placement strategy for deduplication-enabled ZenFS. DeZNS tackles the increased valid data copy overhead during GC in offline deduplication by employing a lightweight CRC32 checksum-based method to predict potential duplicates with minimal performance impact, segregating unique and duplicate data blocks. This segregation reduces valid data migration overhead during GC, while the interruptible GC mechanism ensures that ongoing I/O requests are not delayed during zone resets, maintaining ZenFS performance. Additionally, DeZNS integrates an offline deduplication module that operates on segregated zones. Our extensive evaluation shows that DeZNS reduces valid data migration by 28% compared to baseline ZenFS and by up to 2× compared to naive offline deduplication in microbenchmarks.
- Open Access
- 11 September 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
- Vol. 23 (6) , 1-25
An Analytical Model-based Capacity Planning Approach for Building CSD-based Storage Systems
The data movement in large-scale computing facilities (from compute nodes to data nodes) is categorized as one of the major contributors to high cost and energy utilization. To tackle it, in-storage processing (ISP) within storage devices, such as Solid-State Drives (SSDs), has been explored actively. The introduction of computational storage drives (CSDs) enabled ISP within the same form factor as regular SSDs and made it easy to replace SSDs within traditional compute nodes. With CSDs, host systems can offload various operations such as search, filter, and count. However, commercialized CSDs have different hardware resources and performance characteristics. Thus, it requires careful consideration of hardware, performance, and workload characteristics for building a CSD-based storage system within a compute node. Therefore, storage architects are hesitant to build a storage system based on CSDs as there are no tools to determine the benefits of CSD-based compute nodes to meet the performance requirements compared to traditional nodes based on SSDs. In this work, we proposed an analytical model-based storage capacity planner called CsdPlan for system architects to build performance-effective CSD-based compute nodes. Our model takes into account the performance characteristics of the host system, targeted workloads, and hardware and performance characteristics of CSDs to be deployed and provides optimal configuration based on the number of CSDs for a compute node. Furthermore, CsdPlan estimates and reduces the total cost of ownership (TCO) for building a CSD-based compute node. To evaluate the efficacy of CsdPlan , we selected two commercially available CSDs and four representative big data analysis workloads.
- 11 September 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
- Vol. 23 (6) , 1-25
Open AccessAn Analytical Model-based Capacity Planning Approach for Building CSD-based Storage Systems
The data movement in large-scale computing facilities (from compute nodes to data nodes) is categorized as one of the major contributors to high cost and energy utilization. To tackle it, in-storage processing (ISP) within storage devices, such as Solid-State Drives (SSDs), has been explored actively. The introduction of computational storage drives (CSDs) enabled ISP within the same form factor as regular SSDs and made it easy to replace SSDs within traditional compute nodes. With CSDs, host systems can offload various operations such as search, filter, and count. However, commercialized CSDs have different hardware resources and performance characteristics. Thus, it requires careful consideration of hardware, performance, and workload characteristics for building a CSD-based storage system within a compute node. Therefore, storage architects are hesitant to build a storage system based on CSDs as there are no tools to determine the benefits of CSD-based compute nodes to meet the performance requirements compared to traditional nodes based on SSDs. In this work, we proposed an analytical model-based storage capacity planner called CsdPlan for system architects to build performance-effective CSD-based compute nodes. Our model takes into account the performance characteristics of the host system, targeted workloads, and hardware and performance characteristics of CSDs to be deployed and provides optimal configuration based on the number of CSDs for a compute node. Furthermore, CsdPlan estimates and reduces the total cost of ownership (TCO) for building a CSD-based compute node. To evaluate the efficacy of CsdPlan , we selected two commercially available CSDs and four representative big data analysis workloads.
- Open Access
- 17 October 2023
- Published by Springer Nature in Cluster Computing
- Vol. 27 (2) , 2309
Correction: Scalable NUMA-aware persistent B+-tree for non-volatile memory devices
- 17 October 2023
- Published by Springer Nature in Cluster Computing
- Vol. 27 (2) , 2309
Open AccessCorrection: Scalable NUMA-aware persistent B+-tree for non-volatile memory devices
- Open Access
- 9 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 109-115
A Free-Space Adaptive Runtime Zone-Reset Algorithm for Enhanced ZNS Efficiency
- 9 July 2023
- Published by Association for Computing Machinery (ACM)
- p. 109-115
Open AccessA Free-Space Adaptive Runtime Zone-Reset Algorithm for Enhanced ZNS Efficiency
- Open Access
An Analytical Model-based Capacity Planning Approach for Building CSD-based Storage Systems
The data movement in large-scale computing facilities (from compute nodes to data nodes) is categorized as one of the major contributors to high cost and energy utilization. To tackle it, in-storage processing (ISP) within storage devices, such as Solid-State Drives (SSDs), has been explored actively. The introduction of computational storage drives (CSDs) enabled ISP within the same form factor as regular SSDs and made it easy to replace SSDs within traditional compute nodes. With CSDs, host systems can offload various operations such as search, filter, and count. However, commercialized CSDs have different hardware resources and performance characteristics. Thus, it requires careful consideration of hardware, performance, and workload characteristics for building a CSD-based storage system within a compute node. Therefore, storage architects are hesitant to build a storage system based on CSDs as there are no tools to determine the benefits of CSD-based compute nodes to meet the performance requirements compared to traditional nodes based on SSDs. In this work, we proposed an analytical model-based storage capacity planner called CSDPlan for system architects to build performance-effective CSD-based compute nodes. Our model takes into account the performance characteristics of the host system, targeted workloads, and hardware and performance characteristics of CSDs to be deployed and provides optimal configuration based on the number of CSDs for a compute node. Furthermore, CSDPlan estimates and reduces the total cost of ownership (TCO) for building a CSD-based compute node. To evaluate the efficacy of CSDPlan, we selected two commercially available CSDs and 4 representative big data analysis workloads.
Open AccessAn Analytical Model-based Capacity Planning Approach for Building CSD-based Storage Systems
The data movement in large-scale computing facilities (from compute nodes to data nodes) is categorized as one of the major contributors to high cost and energy utilization. To tackle it, in-storage processing (ISP) within storage devices, such as Solid-State Drives (SSDs), has been explored actively. The introduction of computational storage drives (CSDs) enabled ISP within the same form factor as regular SSDs and made it easy to replace SSDs within traditional compute nodes. With CSDs, host systems can offload various operations such as search, filter, and count. However, commercialized CSDs have different hardware resources and performance characteristics. Thus, it requires careful consideration of hardware, performance, and workload characteristics for building a CSD-based storage system within a compute node. Therefore, storage architects are hesitant to build a storage system based on CSDs as there are no tools to determine the benefits of CSD-based compute nodes to meet the performance requirements compared to traditional nodes based on SSDs. In this work, we proposed an analytical model-based storage capacity planner called CSDPlan for system architects to build performance-effective CSD-based compute nodes. Our model takes into account the performance characteristics of the host system, targeted workloads, and hardware and performance characteristics of CSDs to be deployed and provides optimal configuration based on the number of CSDs for a compute node. Furthermore, CSDPlan estimates and reduces the total cost of ownership (TCO) for building a CSD-based compute node. To evaluate the efficacy of CSDPlan, we selected two commercially available CSDs and 4 representative big data analysis workloads.
- Open Access
- 27 March 2023
- Published by Association for Computing Machinery (ACM)
- p. 1309-1317
MFence: Defending Against Memory Access Interference in a Disaggregated Cloud Memory Platform
- 27 March 2023
- Published by Association for Computing Machinery (ACM)
- p. 1309-1317
Open AccessMFence: Defending Against Memory Access Interference in a Disaggregated Cloud Memory Platform
- 17 November 2022
- Published by Springer Nature in Cluster Computing
- Vol. 26 (5) , 2865-2881
Scalable NUMA-aware persistent B+-tree for non-volatile memory devices
- 17 November 2022
- Published by Springer Nature in Cluster Computing
- Vol. 26 (5) , 2865-2881
Scalable NUMA-aware persistent B+-tree for non-volatile memory devices
- Open Access
- 21 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 114711-114724
Future-Based Persistent Spatial Data Structure for NVM-Based Manycore Machines
R-trees have been popular for their support of multidimensional data and high-performing queries. FBR-tree is the state-of-the-art concurrent variant of the R-tree for Intel DC Persistent Memory (DCPM). However, its adoption on manycore servers is impeded by concurrency limitations due to lengthy, lock-based synchronization, including structure modification operations, split and merge. Additionally, emerging DCPM-based machines are equipped with multiple CPU sockets, forming a non-uniform memory access (NUMA) architecture. FBR-tree’s lack of NUMA-awareness induces further performance overhead from remote memory accesses. In this paper, we propose MPR-tree, a concurrent, NUMA-aware and persistent future-based R-tree for DCPM servers. MPR-tree focuses on insert operations due to their laborious nature. MPR-tree relies on per-thread local future objects and a global R-tree. To introduce NUMA-awareness and minimize remote memory accesses, MPR-tree adopts per-socket dedicated asynchronous evaluate threads to checkpoint future objects to the global R-tree. MPR-tree employs an in-memory hash table to mitigate the read overhead of key searches over the future objects. We implemented MPR-tree atop FBR-tree and evaluated its performance on a server with 40 physical cores for insert and lookup queries, and it showed that MPR-tree outperforms FBR-tree on average by 2× on log10 scale.
- 21 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 114711-114724
Open AccessFuture-Based Persistent Spatial Data Structure for NVM-Based Manycore Machines
R-trees have been popular for their support of multidimensional data and high-performing queries. FBR-tree is the state-of-the-art concurrent variant of the R-tree for Intel DC Persistent Memory (DCPM). However, its adoption on manycore servers is impeded by concurrency limitations due to lengthy, lock-based synchronization, including structure modification operations, split and merge. Additionally, emerging DCPM-based machines are equipped with multiple CPU sockets, forming a non-uniform memory access (NUMA) architecture. FBR-tree’s lack of NUMA-awareness induces further performance overhead from remote memory accesses. In this paper, we propose MPR-tree, a concurrent, NUMA-aware and persistent future-based R-tree for DCPM servers. MPR-tree focuses on insert operations due to their laborious nature. MPR-tree relies on per-thread local future objects and a global R-tree. To introduce NUMA-awareness and minimize remote memory accesses, MPR-tree adopts per-socket dedicated asynchronous evaluate threads to checkpoint future objects to the global R-tree. MPR-tree employs an in-memory hash table to mitigate the read overhead of key searches over the future objects. We implemented MPR-tree atop FBR-tree and evaluated its performance on a server with 40 physical cores for insert and lookup queries, and it showed that MPR-tree outperforms FBR-tree on average by 2× on log10 scale.