Service Based Architecture with Service Mesh Platform In The Context Of 5G Core
Akbarisamani, Mohammadali (2019)
2019
Tietotekniikan DI-ohjelma - Degree Programme in Information Technology
Informaatioteknologian ja viestinnän tiedekunta - Faculty of Information Technology and Communication Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2019-11-26
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-201911035632
https://urn.fi/URN:NBN:fi:tuni-201911035632
Tiivistelmä
A gradual transition from virtual machine based design to new architectures such as service mesh is happening. The service mesh engages with remodeling of application architecture, which ultimately led to a shift in where the service is executed not a shift in what. In this model which is the service mesh, request traffic has an important role in deciding how to re-model the architecture of network.
The objective of this thesis work based on a study item for 3GPP release 16 is Enhanced-SBA. One of the key issues is related to HTTP/2 messaging, which doesn’t have an inbuilt load balancing, overload control or failure recovery mechanisms. One potential solution is in addition to using new radio, consider enhancing with the cloud native grown principles like a service-mesh.
As for the used methods in this thesis, using a service mesh platform being our proposed solution for SBA, research and studies were made between different technologies and Istio was chosen as the most suitable framework for our purpose. It is an open source service mesh platform designed by Google, IBM and Lyft. It layers transparently onto existing distributed applications which lets you run a distributed microservice architecture and runs primarily on Kubernetes.
Istio uses sidecar proxies called Envoys to reduce the complexity of managing microservice deployments by providing a uniform way to secure, connect, and monitor microservices.
The results of this work are presented with conducting an extensive study, which was made on request routing and load balancing principles. Also, applicability to multiple instances of a micro-services (Network Functions), based on different conditions like load, different headers, etc.
The extra latencies introduced by each Envoy was measured under load and different fac-tors affecting the results. However, it must be considered that the measured results depend on the environment where the framework is running, the features enabled, etc. (less than 1 ms in this case).
The objective of this thesis work based on a study item for 3GPP release 16 is Enhanced-SBA. One of the key issues is related to HTTP/2 messaging, which doesn’t have an inbuilt load balancing, overload control or failure recovery mechanisms. One potential solution is in addition to using new radio, consider enhancing with the cloud native grown principles like a service-mesh.
As for the used methods in this thesis, using a service mesh platform being our proposed solution for SBA, research and studies were made between different technologies and Istio was chosen as the most suitable framework for our purpose. It is an open source service mesh platform designed by Google, IBM and Lyft. It layers transparently onto existing distributed applications which lets you run a distributed microservice architecture and runs primarily on Kubernetes.
Istio uses sidecar proxies called Envoys to reduce the complexity of managing microservice deployments by providing a uniform way to secure, connect, and monitor microservices.
The results of this work are presented with conducting an extensive study, which was made on request routing and load balancing principles. Also, applicability to multiple instances of a micro-services (Network Functions), based on different conditions like load, different headers, etc.
The extra latencies introduced by each Envoy was measured under load and different fac-tors affecting the results. However, it must be considered that the measured results depend on the environment where the framework is running, the features enabled, etc. (less than 1 ms in this case).