This article presents an advanced framework for Dynamic Traffic Management (DTM), significantly extending our prior work on minimizing inter-domain traffic costs. The main extension enables cost optimization among multiple sending and multiple receiving domains. The mechanism works for volume and 95thpercentile tariffs. The main paradigm of DTM operation remains unchanged: DTM reacts dynamically on traffic changes on inter-domain links shifting some part of traffic (manageable traffic) from one link to another. This mechanism is based on flow load-balancing, it is centrally controlled by SDN controllers. It requires cooperation between operators receiving and generating traffic. Communication between domain operators is realized by SDN controllers (inter-domain SDN). DTM has been applied to inter-cloud communication for DCs located in different ISP domains. The experiments conducted in the hardware testbed, involving numerous sending and receiving sites, demonstrate that DTM can save up to 36% of the transfer cost.

Despite opening attractive perspectives, the concept of Software Defined Networking raises doubts about the performance and practical feasibility. To contradict these concerns, we propose a deployment-ready system aimed at proactive and periodic optimization of flow paths. The modular system consists of modules responsible for traffic prediction, static optimization, measurements, flow management, and validation of optimization results. To make the system efficient, we resolved several scientific issues and proposed novel and valuable solutions, for example methods for efficient proactive flow management and periodic re-optimization of routing policies. Simultaneously, to make the system production-ready and create a reliable research environment, we provide solutions to several technical obstacles. We validate the proposed system with three network topologies, each with three load levels, following real-life traffic models. We consider Equal-Cost Multi-Path Routing as a baseline. The results indicate that the system allows network operators to handle more traffic (packet loss reduced by up to 30%), improve quality of service (less congested links resulted in even 2.5 times lower latency), and reduce operational expenses (energy consumption lowered by up to 10%).

This article introduces the concept, implementation and evaluation of Dynamic Traffic Management (DTM). This management mechanism minimizes traffic transfer costs on interdomain links. It works for volume and 95 h percentile tariffs. DTM reacts dynamically to traffic changes on inter-domain links shifting some part of traffic (manageable traffic) from one link to another. The management mechanism works in a distributed way. It requires cooperation between operators, namely Internet Service Providers (ISPs), receiving and generating traffic. Communication between ISPs is realized by SDN controllers (inter-domain SDN). DTM has been applied to intercloud communication for Data Centers (DCs) located in different ISP domains. Moreover, it is cloud-agnostic, i.e., DTM does not engage Cloud Service Providers (CSPs) into management procedures. Evaluation performed in the testbed shows that DTM is capable of decreasing ISP costs of transit traffic up to 29%.

This paper is devoted to analysis of QoS parameters of WebRTC traffic presented for centralized videoconferencing system used for collaborative work. This system consists of a videoconference application, built according to the WebRTC architecture, and a telemetric system, built as the IoT environment. Tests were carried out for three locations of the conference bridge: in private cloud (the OpenStack cloud), in public cloud (the AWS cloud) and in local network. Results show, that conference bridge in a cloud is a good solution for the WebRTC, and additional transmission of telemetric data do not affect the QoS parameters of WebRTC's media stream.