Modellgestützte Analyse von zukünftigen Natzarchitekturen und Komponenten
|Duration:||3 years, 01.07.2018-30.06.2020|
|Partner:||TUM Chair of Network Architectures and Services (I8)|
|Contact:||Raphael Durner (email@example.com)|
It is foreseen, that new paradigms in computer networking like Network Function Virtualization (NFV), Network Virtualization (NV) and Software Defined Networking (SDN), will increase the flexibility and openness of modern communication infrastructure. The new approaches will enable virtualization of network functions (NFV), network slicing (NV) and a (logically) centralized control of the network. Therefore, functionality can be decoupled from expensive specialized hardware towards virtualized commodity hardware, e.g. in data centers. This leads to an ongoing separation of network services from the underlying hardware.
Legacy networks are commonly planned and implemented based on vendor recommendations and experience. With the help of traffic engineering techniques, it is possible to steer aggregated traffic flows efficiently in the network and to influence the properties of certain service classes. Traffic engineering can be done in a static way, based on previously collected traffic statistics, or dynamically, based on monitoring in the network. SDN enables novel approaches to analyze and control the traffic in the network, however, requiring a deep understanding of the underlying hardware. While the packet processing pipeline of well-designed, specialized, hardware is specified in detail, virtualized software on commodity hardware is much more difficult to tackle. On the other hand, the additional degrees of freedom enabled by virtualizing network functions (NFV), such as custom packet processing pipelines and dynamic placement, create new possibilities for optimization.
Goal of the project is to develop models for the planning and dynamic optimization of virtualized network functions and network elements. To that end, we first have to identify and model each step in the packet processing pipeline separately. Afterwards, we combine the synthesized models to characterize the performance of the whole virtualized processing pipeline. This will allow us to accurately predict the performance of virtualized forwarding paths in the network and based on that, enables us to implement traffic engineering and other traffic flow optimization techniques. In the evaluation we take in account dedicated hardware switches, like OpenFlow-enabled switches, but also software-based switches. Furthermore, in addition to the performance characteristics, we will evaluate the cost factors for operating the devices (e.g. energy consumption of the network elements). The outcomes of the project will provide a deep understanding of the characteristics of the different processing pipelines in SDN-based networks. This knowledge will allow optimizations on the scope of a device (e.g.routing algorithms), but will also facilitate the design of novel network elements combining software and hardware processing capabilities and improved approaches for network design as a whole.
- Novel Models for softwarized and virtualized packet processing systems
- Models on component, node and network level
- Measurements to derive and support new models