Titulo Estágio
What are the most dangerous failures in Microservices?
Áreas de especialidade
Engenharia de Software
Comunicações, Serviços e Infraestruturas
Local do Estágio
SSE-CISUC
Enquadramento
Microservices adoption are an increasing trend. The idea is that complex systems can be decomposed in smaller fine-grained services that communicate through a well-defined, lightweight mechanism. This allows these services to scale independently.
This architecture model suits well the recent cloud trends and when used in conjunction with containers technologies, it may take advantage of all the possibilities associated. For instance, it is quite fast to deploy new containers to adjust to the demand, and with kubernetes this scaling can be even easier to automate.
Containerization can even help with availability, as failures can be isolated in microservices, and the failing microservices can be easily stopped and replaced.
However, this is definitely not silver bullet, as there are several failure modes that are much harder to detect, such as silent data corruptions, undetected performance degradation or incorrect results.
In fact, some of these failure modes may be masked or even exacerbated by the distributed nature of the architectural model, due to the impact of the network and also data synchronization.
Objetivo
This project aims at understanding which are the most dangerous types of failures for the existent best practices for microservices failure recovery. For this, fault injection techniques will be used to introduce faults that will lead to failures of one microservice of the system, in order to evaluate different applications that were designed following the most recommended best practices for microservices architectures deployment. With this, it will be possible to observe how the existent systems deal with different types of failures and understand which ones are the most concerning for the current practices in microservices.
The student will gain extensive practical experience with technologies that are increasingly being adopted.
Plano de Trabalhos - Semestre 1
[10/09/2018 to 31/10/2018] State of the art analysis.
[15/10/2018 to 30/11/2018] Systematization of the best practices for fault tolerance and failure recovery in microservices. Analysis of the most relevant failures.
[01/11/2018 to 31/12/2018] Definition of the experimental approach. Definition of the systems to be used during the experiments.
[15/11/2018 to 21/01/2019] Write the Dissertation Plan.
Plano de Trabalhos - Semestre 2
[01/02/2019 to 15/04/2019] Implementation of the tools to support the experimental approach.
[20/03/2019 to 30/04/2019] Experimentation and validation.
[15/04/2018 to 31/05/2019] Write a scientific publication.
[15/05/2019 to 01/07/2019] Write the thesis.
Condições
The work is to be executed at the laboratories of the CISUC’s Software and Systems Engineering Group. A work place will be provided as well as the required computational resources.
Observações
This work is part of the contribution of the U. Coimbra to the H2020 Project ATMOSPHERE, which increases the impact and visibility of the work and its results.
Orientador
Nuno Antunes
nmsa@dei.uc.pt 📩