Micro-architectural Analysis of In-memory OLTP

  • Utku Sirin, Pinar Tözün, Danica Porobic, Anastasia Ailamaki
  • June 2016, ACM (Association for Computing Machinery)
  • DOI: 10.1145/2882903.2882916

The Efficiency of Transaction Processing Systems at Hardware Level

Photo by Slejven Djurakovic on Unsplash

Photo by Slejven Djurakovic on Unsplash

What is it about?

Online Transaction Processing, i.e., OLTP, is a multi-billion dollar industry. As a result, every single improvement over the OLTP system performance has a significant monetary return. How do we improve OLTP system performance? How can we judge whether the OLTP system is delivering as high performance as it can? Does the OLTP system is at its limit in terms of the delivered performance? Interesting questions, whose answers are in our work.

Why is it important?

OLTP system performance is one of the most important concerns for system designers, and hardware is one of the most important resources for delivering high performance.

Perspectives

Utku Sirin
Harvard University

This work considers a wide range of systems running on the same hardware with the same configuration. Therefore, its results are representative of a wide range of OLTP and alike workloads for OLTP and alike systems.

Read Publication

http://dx.doi.org/10.1145/2882903.2882916

The following have contributed to this page: Utku Sirin