Latency Optimized Microservice Architecture for Privacy Preserving Computing

Abstract

Recent developments in microservices architecture and building have lead to the advent of unikernels, a specialized operating system kernel coupled with, and executing only, a single application.

This thesis presents PPCE a distributed system utilising a microservices architecture based on unikernels, created to enable privacy-preserving computing for users, classes of users, and more importantly; doing so with acceptable latency.

We built an initial version of \ac{ppce}, and through experimentation, we measured the latency to be 63.15 ms. Then we devised two optimized approaches to reduce the user-perceived latency of the initial version.

After conducting experiments on the two optimized versions of PPCE, we discussed the applicability of the two approaches and compared them to related work. Showing that our hybrid version of PPCE is able to implement an architecture enabling privacy preserving computing with only 29.53 ms user-perceived latency.