D
Dipti Shankar
Researcher at Ohio State University
Publications - 24
Citations - 413
Dipti Shankar is an academic researcher from Ohio State University. The author has contributed to research in topics: Remote direct memory access & InfiniBand. The author has an hindex of 10, co-authored 24 publications receiving 368 citations.
Papers
More filters
Proceedings ArticleDOI
Accelerating Spark with RDMA for Big Data Processing: Early Experiences
TL;DR: This paper proposes a high-performance RDMA-based design for accelerating Spark with RDMA for Big Data processing and adopts a plug-in-based approach that can make the design to be easily integrated with newer Spark releases.
Proceedings ArticleDOI
Triple-H: a hybrid approach to accelerate HDFS on HPC clusters with heterogeneous storage architecture
TL;DR: This paper presents a hybrid design (Triple-H) that can minimize the I/O bottlenecks in HDFS and ensure efficient utilization of the heterogeneous storage devices available on HPC clusters and improves the write and read throughputs of HDFS.
Proceedings ArticleDOI
High-performance design of apache spark with RDMA and its benefits on various workloads
TL;DR: The RDMA-based Spark design is implemented as a pluggable module and it does not change any Spark APIs, which means that it can be combined with other existing enhanced designs for Apache Spark and Hadoop proposed in the community.
Proceedings ArticleDOI
Accelerating I/O Performance of Big Data Analytics on HPC Clusters through RDMA-Based Key-Value Store
TL;DR: This study designs a burst buffer system using RDMA-based Mem cached and presents three schemes to integrate HDFS with Lustre through this buffer layer, considering different aspects of I/O, data-locality, and fault-tolerance.
Proceedings ArticleDOI
Performance characterization and acceleration of in-memory file systems for Hadoop and Spark applications on HPC clusters
TL;DR: This paper characterizes two file systems in literature, Tachyon and Triple-H, that support in-memory and heterogeneous storage, and discusses the impacts of these two architectures on the performance and fault tolerance of Hadoop MapReduce and Spark applications.