COMMUNITY

BOARD

News

Outstanding Presentation Paper Award at the Korea Software Congress 2024 (KSC 2024) - DISCOS Laboratory -

Author
College of Software Convergence
Date
2025-02-15
View
31

Files

Outstanding Presentation Paper Award at the Korea Software Congress 2024 (KSC 2024) - DISCOS Laboratory -



At the Korea Software Congress 2024 (KSC 2024), held over three days from 18 to 20 December, the paper 'Design of a Multi-CSD-based Document Database for Accelerating Query Throughput', presented by master's student Kim Ki-hwan (supervisor: Kim Young-jae), received the Outstanding Presentation Paper Award.

 

The paper is the result of research written jointly by Yoo Hyung-seon (undergraduate researcher), Kim Ki-hwan (master's), Yoo Jung-hyun (master's) and Byun Hong-su (doctoral) under the supervision of Professor Kim Young-jae.



▲ From left: Yoo Hyung-seon (undergraduate researcher), Kim Ki-hwan (master's), Yoo Jung-hyun (master's), Byun Hong-su (doctoral)

▲ Certificate of the Outstanding Presentation Paper Award, Korea Software Congress 2024 (KSC 2024)



Existing distributed databases suffer limits to scale-out scalability because of constraints such as network bottlenecks and the overhead of maintaining consistency. This research aims to overcome those scale-out limits and improve query processing performance through a multi-CSD-based document database design that exploits computational storage drives (CSD) — actively researched in recent years — and features within CouchDB, a representative document database. By presenting optimization of shard count allocation within CSD devices and a compute-layout-aware data distribution method, it improved the query processing performance of the document database.

Existing distributed databases suffer limits to scale-out scalability because of constraints such as network bottlenecks and the overhead of maintaining consistency. This research aims to overcome those scale-out limits and improve query processing performance by designing, for the first time, a multi-CSD-based document database using computational storage drives (CSD) — actively researched in recent years — together with the cluster mode feature of CouchDB, a representative document database. To that end it proposes optimization of shard count allocation within CSD devices and a compute-layout-aware data distribution technique, improving the query processing performance of the document database.

The Korea Software Congress 2024 (KSC 2024) is an information science and technology conference at which the latest research results across all fields of information technology — AI and deep learning, computer systems, databases and more — are presented.