Seminar • RAMP: RDMA Migration Platform

Babar Naveed Memon, Master’s candidate
David R. Cheriton School of Computer Science

Remote Direct Memory Access (RDMA) can be used to implement a shared storage abstraction or a shared nothing abstraction for distributed applications. We argue that the shared storage abstraction is an overkill for loosely coupled applications and that the shared nothing abstraction does not leverage all the benefits of RDMA.

In this thesis, we propose an alternative abstraction for such applications using a shared-on-demand architecture, and present the RDMA Migration Platform (RAMP). RAMP is a lightweight coordination service for building loosely coupled distributed applications. This thesis describes the RAMP system, its programming model and operations, and evaluates RAMP using micro-benchmarks. Furthermore, we illustrate RAMPs load balancing capabilities by simulating a real application workload in a loosely coupled distributed application.

Bio: Babar Naveed Memon is a graduate student from the Data Systems Group with strong interest in building distributed systems and databases. His is research primarily involved in leveraging RDMA for cloud-based applications. After graduation he intends to work at SAP in Waterloo building distributed databases and actively participating in research in the database and systems community. He is also a big sports fan and food enthusiast.