A database scale-out solution for emerging write-intensive commercial workloads

  • Authors:
  • Yi Ge;Chen Wang;Xiaowei Shen;Honesty Young

  • Affiliations:
  • IBM China Research Lab;IBM China Research Lab;IBM China Research Lab;IBM China Research Lab

  • Venue:
  • ACM SIGOPS Operating Systems Review
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Among the increasing number of online businesses, a series of write-intensive commercial workloads are emerging recently on the Internet. These workloads generate many more write transactions on the backend database than read transactions. Most of such workloads require the database to handle high-volume write transactions in real-time. Based on the observation on the workloads, this paper proposes a multi-tier database scale-out architecture with the write caching technology. Our preliminary result shows that the database scale-out architecture can handle extremely high-volume write transactions with excellent scalability.