Seminar • Systems and Networking • An Efficient Key-Value Store for Hardware Disaggregation Architecture

Friday, November 3, 2023 2:30 pm - 3:00 pm EDT (GMT -04:00)

Please note: This half-hour seminar will take place in DC 1304.

Ahmed Alquraan, PhD candidate
David R. Cheriton School of Computer Science

Supervisor: Professor Samer Al-Kiswany

Hardware disaggregation is the practice of separating various hardware components of a computing system, such as CPU, memory, and storage, into distinct, interoperable modules. This approach offers benefits like flexibility, improved resource utilization, improved fault tolerance, and cost savings by enabling components to be upgraded or replaced independently. On the other hand, hardware disaggregation imposes several design challenges including network and data transfer overhead. Current distributed key-value stores are not designed with hardware disaggregation in mind because they assume that a single machine can host all data structure needed for the system to operate as well as that most memory accesses are local to the CPU, which is not the case with hardware disaggregation.

In this project, we are building a new key-value store that addresses the aforementioned shortcomings and achieves better performance and resource utilization with hardware disaggregation architecture.


Bio: Ahmed Alquraan is a 5th-year PhD student working with Samer Al-Kiswany. His work focuses on utilizing new data center technologies to build efficient and reliable data stores.


Please note: This is the second of two back-to-back Systems and Networking seminars.