Teasing Out Causal Relationships in Distributed Systems by Adding Delays to gRPCs

May 10, 2024
10:30am
Cummings 302
Speaker: Mona Ma - Master's Defense
Host: Raja Sambasivan

Abstract

Master's Defense:

Microservices are fundamental to the architecture of contemporary distributed systems, which handle requests in sequential, concurrent, and parallel manners to facilitate scalability. To manage concurrency, various programming languages implement deferred computing techniques such as futures and promises. However, the causal relationships and dependencies among microservices often remain unclear. This thesis explores the potential of introducing delays into gRPC calls to reveal these causal relationships within distributed systems. The results aim to provide developers and researchers with valuable insights into enhancing system efficiency by prioritizing optimizations for critical components.

Areas of research: Cloud computing, distributed systems