Software-defined networking strategies for efficient next-generation applications

Abstract

Next-generation Internet Applications (NIA), such as smart-cities and multi-media applications like virtual reality, provide the technological foundation and capabilities necessitated for digital transformation in various economic sectors. With the widespread adoption of these applications, network requirements have become increasingly demanding e.g., they require ultra-low delay for rapid decision-making and high throughput for transferring large amounts of data. In this thesis, we focus on the development of efficient network strategies in order to improve the performance of two specific types of NIA: IoT and cloud-native applications.

We focus on two primary network environments that require adaptability to fulfill the needs of the aforementioned NIA: (i) large-scale wireless-based IoT deployments (e.g., smart cities) and (ii) cloud environments that host microservice-based applications with heterogeneous resource demands. In this context, we investigate the potential advantages derived from a synergy of cloud computing, Software-Defined Networking (SDN), Named-Data Networking (NDN), and the microservices paradigm in enhancing NIA performance through efficient traffic steering. Also, we argue that the centralized control feature of SDNs is an appropriate solution for integrating these technologies and creating a highly adaptive network ecosystem.

The research is concentrated around three key areas: (i) improving IoT applications' adaptability to Wireless Mesh Network (WMN) to ensure consistent and reliable performance; (ii) implementing microservices adaptive load balancing, which implements efficient resource allocation and utilization, enabling applications to handle a larger number of simultaneous requests without performance compromise; and (iii) predicting the impact of individual microservices on network and processing resources, thereby support the decisions of resource allocation control mechanisms.

In order to tackle the aforementioned challenges, this research focuses on the development and evaluation of realistic applications and SDN network environments over WMN, including as well as real indoor and outdoor testbeds. The objective is to promote a framework where NIA can exhibit greater responsiveness and resource efficiency by adapting to the dynamic WMNs and cloud resource requirements.