The main project objective is to define and test an Ethernet-based architecture able to guarantee the communication services of a satellite launcher. The current available solutions are not suited to be applied in launcher networks because either they are not able to guarantee deterministic and predictable behaviour or they are too costly as the TTEthernet standard.
I propose a low cost solution able to overcome the disadvantages of the Ethernet standard that is the lack of guaranteeing Quality of Service.
The proposed solutions will be based on the following architectural/technological choices:
-use of Common-Off-The-Shelf (COTS) Ethernet technology as much as possible with the addition of the minimum number of functional components for the support of real-time traffic;
-use of traditional Ethernet switches, full duplex transmission and simple network topology in order to reduce the number of both nodes and ports;
-implementation of a real-time layer bypassing the TCP/UDP/IP protocol stack to avoid potential delays introduced by this protocols;
-definition of a layer for the scheduling of messages so as to guarantee the QoS requested by the Guide, Navigation and Control (GNC) and telemetry applications.
I will start from the Flexible Time Triggered Ethernet (FTTE) standard and modify it so as to develop an ad hoc solution for launcher network.
The project is organized in three Work Packages. The WP1 is devoted to identify the networking requirements and study the real time Ethernet networks based on COTS technology. I identify and investigate FFTE-based configurations for the launcher communication system in WP2; a message scheduling algorithm is also defined in this WP. I verify the effectiveness of the proposed solution in WP3 where I develop an experimental test-bed; after developing open source software I perform measure to verify: i) the correct operation mode of the proposed solution; ii) its ability in guaranteeing communications with hard delay.
The FFT Ethernet-based network architecture for launchers has the possibility to guarantee high bandwidth, not less than 100 Mb/s and to guarantee Quality of Service to safety-critical applications. That means a 100 times increases in supported bandwidth with respect to traditional solutions as bus 1553B. The solution is costly because it is inherited by the Ethernet technology and re-use as much as possible Ethernet COTS devices.
It also allows for the sharing of the network infrastructures between safety-critical terminals (i.e. OBC, NU, EC,¿) and non-critical ones (i.e. VC, TRU). The network switches and topology, in fact, assure the proper reliability for the different users without having to implement a fail-silent interface on the non-critical systems. This is achieved by silencing a bubbling idiot terminal at switch-level (denying its transmissions on the virtual links) instead of at terminal level as it is done, for example, on the 1553B bus (no bus chattering is requested on 1553 remote terminals). This means that the non-critical end systems can be connected to the same communication system of critical ones without having to physically segregate the two sub-systems as it is done in today¿s launcher networks. This implies a great gain in terms of system complexity and harness limitation with, as consequence, a limitation of weight, cables and connectors, tests to be performed at system level and, therefore, a cost reduction both on development and in production phases.
I will compare the performance of the proposed solution with respect to traditional solutions as TTEthernet. In order to carry out this comparison I will develop a frame scheduling algorithm that avoids frame contentions on the network links.
I will also develop an experimental test-bed in which I will implement the proposed solution and test it in a real traffic scenario. I will put at disposal open source codes that implement the both scheduling algorithm and the solution based on FFT Ethernet protocol.