SANDY – Synchronised Access Networks for Distributed real-time sYstems

Software architecture for optimising time-sensitive data transmission via the Internet.

The goal of the project “Synchronised Access Networks for Distributed real-time sYstems” (SANDY) is the development of an IP overlay architecture that offers deterministic low delay guarantees for real-time traffic flows. The goal is to offer a wide-area networking solution for the coupling of embedded systems operating under real-time networking constraints.

The major innovations of SANDY are made in the access networks where the concept of a time-triggered synchronisation of traffic flows is introduced at the IP layer. The figure below shows 3 synchronized flows.

To enable low and deterministic delay bounds, distinct strategies are applied in the different regions of a wide area IP network. The distinction is made between access networks, core networks, and inter-domain areas as shown in the figure below.

It is shown that a worst case delay computation is a feasible approach in the core network. In the inter-domain area, the concept of the Differentiated Service’s bandwidth broker is extended to include the worst case delays that may occur within network domains. The major innovations of SANDY are made in the access networks. The concept of a time-triggered synchronisation of traffic flows is introduced at the IP layer. When an end-system’s request for a real-time flow is accepted by the network resource manager, the sender is allocated a set of time slots at which it must emit its packets. The allocation of time slots is enforced in a way such that at most one IP packet arrives within a time slot at the output port of each ingress edge device. Additionally, the time slot allocation between the ingress and egress edge devices are synchronised. This strategy enables a conflict free flow of IP packets across the access networks of the wide area connection.

A major focus will be put on the development of scalable resource management algorithms to enable a high resource utilisation and thus a practically feasible solution.

 
How to find us
Salzburg Research Forschungsgesellschaft
Jakob Haringer Straße 5/3
5020 Salzburg, Austria