Kafkalytics

Title

Kafkalytics: Anti-pattern detection in a Kafka message bus

 Author

Slag, R.

Contributor

Van Deursen, A. (mentor)

Faculty

Electrical Engineering, Mathematics and Computer Science

 Department

Software Engineering

 Date

2016-06-21

 Abstract

Recently microservices have emerged as a new architectural pattern which promises many advantages. Services are modeled along business entities, which should result in a flexible system. Apart from that the pattern promises better fault resilience against outages and better performance regarding scalability. In this paper we explore the differences between several architectural styles where we focus on microservices. Additionally we discuss the actual usage of the architures in practice, based on interviews with industry experts. Talks with these experts indicated several problems regarding communication between services. We conclude by identifying several antipatterns when using a Kafka message bus and present a tool Kafkalytics which can detect these patterns. Subsequently Kafkalytics has been implemented in a live production environment where it was evaluated as a tool with low overhead which is able to detect various antipatterns in a live production environment. It allows to be gradually implemented in a complete system and leverages existing infrastructure systems.

 To reference this document use:

http://resolver.tudelft.nl/uuid:8cf81c2f-6ccb-4136-a6bc-51dfc8ffcab8

 Part of collection

Student theses

 Document type

master thesis

 Rights

(c) 2016 Slag, R.
Files
PDF thesis.pdf 453.44 KB
Close viewer