An In-Switch Architecture for Low-Latency Microservices

Loading...
Thumbnail Image

Date

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

In recent time, there is has been a movement away from standard monolithic architecture in cloud and web services towards what is known as a microservice architecture. Microservice architecture decomposes the previous monolithic architecture into multiple independent services called "microservices". Examples of applications that use a microservice architecture include Netflix and Amazon. These applications typically send large numbers of microservice requests, which go through the OSI network layers to establish a client server connection. This trend towards microservices has developed interest by other researchers to make improvements in this field, due to the growing reliance importance on such architectures by consumers. There have been studies regarding the security of these microservices, performance analysis of various applications, and the use of these microservice applications in cloud technology. Any improvements in the speed, security, or organization of such network architecture would be very beneficial of these popular API's, and their user base. This project's objective is to investigate the potential of moving some of the processing that is done for these microservices within a network switch, and as a result the performance at the application level, by alleviating network communication. We formulate a high-level design for an in-switch architecture for low-latency microservice leveraging existing programmable-switches support. We investigate the implementation of NetCache as a microservice in our model and predict a significant latency reduction and subsequent performance increase.

Description

Keywords

Microservice, In-Switch Architecture, In-Switch, Cloud Computing, NetCache

Citation