Figure 2 - uploaded by Ahmed Amari
Content may be subject to copyright.
Source publication
![Figure 1.2: Avionic network architecture [3]](profile/Ahmed-Amari/publication/326007511/figure/fig1/AS:641968925335552@1530068727111/Avionic-network-architecture-3_Q320.jpg)
![Figure 1.4: End-to-end message transmission in the AFDX [4]](profile/Ahmed-Amari/publication/326007511/figure/fig3/AS:641968925339648@1530068727188/End-to-end-message-transmission-in-the-AFDX-4_Q320.jpg)
The inherent complexity and bandwidth requirement of avionics communication architectures are increasing due to the growing number of interconnected end-systems and the expansion of exchanged data. The Avionics Full Duplex Switched Ethernet (AFDX) has been introduced to provide high-speed communication (100Mbps) for new generation aircraft. However...
Contexts in source publication
Context 1
... 1000BASE-T uses a continuous signaling system. Even between the transmission of two frames or at the lack of transmission, it continues to transmit a set of control symbols to improve the synchronization. Fig. 2.4 shows an example of a multilevel signaling in ...
Context 2
... correct errors. If the PHY can not correct the errors and detects that the quality of the communication flow is degraded, then the 1000BASE-T switches from 1000 Mbps mode to a lower mode (100 Mbps or 10 Mbps) to better control the communication and generates more redundancies to correct the data. The high error rate, i.e., equals or near to 1 in Fig. 2.5, corresponds to a loss of connection, where both PHYs perform the auto-negotiation mechanism to negotiate a transmission mode able to resist to the interferences, i.e. reduces the data rate and increases the redundancy. Hence, the disappearance of losses can have two explanations: 1) the two PHYs were able to negotiate an operating ...
Context 3
... Level Ring (DLR) [48,49] and PowerLink [50]. Finally, the third category of solutions modifies the standard implementation, by introducing a set of hardware modifications, to guarantee a better real-time behaviour. A different classification is introduced herein to distinguish the main RTE solutions from an avionics perspective, as shown in Fig. 2.7. Four classes of RTE solutions supporting ring topology have been ...
Context 4
... by the EtherCAT Technology Group (ETG). It implements a master/ slave mechanism on top of Fast Ethernet (100Mbps). The main particularity of EtherCAT is the on-the-fly forwarding technique, which allows the slaves to insert the requested data in a standard Ethernet frame crossing the couplers step by step. As shown in the first topology of Fig. 2.9, EtherCAT is a line network. Thanks to the Full-Duplex links, frames are sent by the master until the last slave, which sends back the frame to the master in the opposite direction to form a virtual ring on a line topology. It is worth noting that this technique requires a specific implementation within the slaves, but allows at the ...
Context 5
... noting that this technique requires a specific implementation within the slaves, but allows at the same time collecting data from several slaves to be transmitted within the same frame. Therefore, this technique allows reducing the overhead of EtherCAT to one header for many collected data, instead of one header per data in classic Ethernet (see Fig. 2.8). EtherCAT ensures a great flexibility by adding and reconfiguring the nodes on-the-fly, thanks to the "hot connect" mechanism. Furthermore, to guarantee the reliability require- ments, EtherCAT supports the master redundancy due to the hot standby method, and imple- ments a dynamic redundancy solution based on a ring topology. In ...
Context 6
... to reduce the cycle time, PROFINET/IRT implements a slipstream method to transmit data, which consists in sending the packets following the physical order of the nodes from the master point of view: the first packet is for the farthest node and the last packet is for the nearest node. This method decreases inherently the communication latencies. Fig. 2.10 shows a temporal diagram of a PROFINET communication with the slipstream effect. The cycle time is the time between the transmission of the first bit of the first packet by the master until the reception of the last bit of the last packet by the first node. Furthermore, Profinet/IRT supports reliability features through implementing ...
Context 7
... III communications are based on cycles, where each cycle consists of two logical communication channels: the RT channel for real-time communications and IP channel for non-real time communications. As shown in Fig. 2.11, each cycle is initiated by the master and consists in sending up to four Master Data Telegrams (MDT) and four Acknowledge Telegrams (AT) during the RT channel, and then the IP channel where slaves can send their non-real-time data in standard frames. The number and duration of MDTS and ATs are configured during the initialization ...
Context 8
... (Very High Performance Automation Bus System) is an academic RTE solution proposed in 2013 by [56]. Unlike many RTE protocol using the standard Ethernet frame to convey their data, VABS has its own frame format and its own data link layer protocol, as shown in Fig 2.12. This choice limits VABS compatibility with standard Ethernet devices. ...
Context 9
... choice limits VABS compatibility with standard Ethernet devices. Figure 2.12: VABS frame structure VABS is based on the master/slave mechanism to convey synchronous traffic and a token passing for asynchronous traffic. ...
Context 10
... links propagation delays (round trip); Figure 2.15: EtherCAT spatio-temporal diagram ...
Context 11
... model the Profinet IRT cycle time, we consider the slipstreamimg effect [62,58,57,61], which consists in sending frames in the same physical order of the destinations from the master. Fig. 2.10 shows a spatio-temporal diagram of Profinet. The slipstreaming effect is beneficial and considered only when it is positive, which means τ ≥ δ + l , i.e., α ≥ 0. Consequently, the Profinet IRT MCT ...
Context 12
... to assess the availability level of the different RTE solutions, the maximum recovery time is shown in Fig. 2.21. The results confirm again the first qualitative conclusions in Section 2.2.5, where EtherCAT and Profinet/IRT have almost similar availability levels, which are much better than the one offered by ...
Context 13
... es− > mar k() × nod e_ f ai l ur e_r at e As explained in Section 5.2.1 and shown in Fig.5.2, the failure rate of a T-AeroRing is the sum of failure rates of its entities, i.e., the Node Core, the three interfaces, the power supply and the connecting card. ...
Context 14
... has been defined by Beckhoff GmbH and supported by the EtherCAT Technology Group (ETG). It implements a master/ slave mechanism on top of Fast Ethernet (100Mbps). The main particularity of EtherCAT is the on-the-fly forwarding technique, which allows the slaves to insert the requested data in a standard Ethernet frame crossing the couplers step by step. As shown in the first topology of Fig. 2.9, EtherCAT is a line network. Thanks to the Full-Duplex links, frames are sent by the master until the last slave, which sends back the frame to the master in the opposite direction to form a virtual ring on a line topology. It is worth noting that this technique requires a specific implementation within the slaves, but allows at the same time collecting data from several slaves to be transmitted within the same frame. Therefore, this technique allows reducing the overhead of EtherCAT to one header for many collected data, instead of one header per data in classic Ethernet (see Fig. 2.8). EtherCAT ensures a great flexibility by adding and reconfiguring the nodes on-the-fly, thanks to the "hot connect" mechanism. Furthermore, to guarantee the reliability require- ments, EtherCAT supports the master redundancy due to the hot standby method, and imple- ments a dynamic redundancy solution based on a ring topology. In the case of a link or node failure, first, the slave detecting the failure returns immediately the EtherCAT frame to the master to avoid losing the communication with the rest of the nodes. Afterwards, the master activates its ports and sends the frame on both to be received by all slaves. Furthermore, the master can determine the failure location through analyzing the slaves error ...
Context 15
... has been defined by Beckhoff GmbH and supported by the EtherCAT Technology Group (ETG). It implements a master/ slave mechanism on top of Fast Ethernet (100Mbps). The main particularity of EtherCAT is the on-the-fly forwarding technique, which allows the slaves to insert the requested data in a standard Ethernet frame crossing the couplers step by step. As shown in the first topology of Fig. 2.9, EtherCAT is a line network. Thanks to the Full-Duplex links, frames are sent by the master until the last slave, which sends back the frame to the master in the opposite direction to form a virtual ring on a line topology. It is worth noting that this technique requires a specific implementation within the slaves, but allows at the same time collecting data from several slaves to be transmitted within the same frame. Therefore, this technique allows reducing the overhead of EtherCAT to one header for many collected data, instead of one header per data in classic Ethernet (see Fig. 2.8). EtherCAT ensures a great flexibility by adding and reconfiguring the nodes on-the-fly, thanks to the "hot connect" mechanism. Furthermore, to guarantee the reliability require- ments, EtherCAT supports the master redundancy due to the hot standby method, and imple- ments a dynamic redundancy solution based on a ring topology. In the case of a link or node failure, first, the slave detecting the failure returns immediately the EtherCAT frame to the master to avoid losing the communication with the rest of the nodes. Afterwards, the master activates its ports and sends the frame on both to be received by all slaves. Furthermore, the master can determine the failure location through analyzing the slaves error ...
Context 16
... 1000BASE-T uses a continuous signaling system. Even between the transmission of two frames or at the lack of transmission, it continues to transmit a set of control symbols to improve the synchronization. Fig. 2.4 shows an example of a multilevel signaling in ...
Context 18
... results can be explained due to the code groups used in the Gigabit Ethernet to control the flow of communication and correct errors. If the PHY can not correct the errors and detects that the quality of the communication flow is degraded, then the 1000BASE-T switches from 1000 Mbps mode to a lower mode (100 Mbps or 10 Mbps) to better control the communication and generates more redundancies to correct the data. The high error rate, i.e., equals or near to 1 in Fig. 2.5, corresponds to a loss of connection, where both PHYs perform the auto-negotiation mechanism to negotiate a transmission mode able to resist to the interferences, i.e. reduces the data rate and increases the redundancy. Hence, the disappearance of losses can have two explanations: 1) the two PHYs were able to negotiate an operating mode able to resist to the interferences; 2) the generated interferences were decorrelated from the transmitted signal, i.e., did not affect the ...
Context 19
... (Isochronous Real-Time) is an extended version of PROFINET, which supports real time communications on top of Fast Ethernet (100Mbps). It is a master/ slave network, based on cyclic communication handling two communication channel types: isochronous and asynchronous. These channels are used by slaves to transmit real-time and non real- time data, respectively. The data is relayed using the Cut-through mechanism to reduce the processing time. The first communication channel, i.e. isochronous, is called time scheduled communication, where all packets transmissions are scheduled during the initialization phase. It is worth noting that the isochronous channel requires an accurate synchronization protocol to guarantee the packet transmissions according to a predefined schedule. The second channel is called, SRT channel (Soft Real-Time channel), used to satisfy the real-time automation constraints. It is based directly on Ethernet (Layer 2) which reduces the processing time within the communication stack. The third communication channel is reserved to non-real-time TCP/UDP/IP packets without temporal constraints. Furthermore, in order to reduce the cycle time, PROFINET/IRT implements a slipstream method to transmit data, which consists in sending the packets following the physical order of the nodes from the master point of view: the first packet is for the farthest node and the last packet is for the nearest node. This method decreases inherently the communication latencies. Fig. 2.10 shows a temporal diagram of a PROFINET communication with the slipstream effect. The cycle time is the time between the transmission of the first bit of the first packet by the master until the reception of the last bit of the last packet by the first node. Furthermore, Profinet/IRT supports reliability features through implementing the MRP [37], based on a ring ...
Context 20
... to assess the availability level of the different RTE solutions, the maximum recovery time is shown in Fig. 2.21. The results confirm again the first qualitative conclusions in Section 2.2.5, where EtherCAT and Profinet/IRT have almost similar availability levels, which are much better than the one offered by ...
Context 21
... III communications are based on cycles, where each cycle consists of two logical communication channels: the RT channel for real-time communications and IP channel for non-real time communications. As shown in Fig. 2.11, each cycle is initiated by the master and consists in sending up to four Master Data Telegrams (MDT) and four Acknowledge Telegrams (AT) during the RT channel, and then the IP channel where slaves can send their non-real-time data in standard frames. The number and duration of MDTS and ATs are configured during the initialization ...
Context 22
... es− > mar k() × nod e_ f ai l ur e_r at e As explained in Section 5.2.1 and shown in Fig.5.2, the failure rate of a T-AeroRing is the sum of failure rates of its entities, i.e., the Node Core, the three interfaces, the power supply and the connecting ...
Context 23
... (Very High Performance Automation Bus System) is an academic RTE solution proposed in 2013 by [56]. Unlike many RTE protocol using the standard Ethernet frame to convey their data, VABS has its own frame format and its own data link layer protocol, as shown in Fig 2.12. This choice limits VABS compatibility with standard Ethernet devices. Figure 2.12: VABS frame structure VABS is based on the master/slave mechanism to convey synchronous traffic and a token passing for asynchronous traffic. As EtherCAT and SERCOS III, VABS supports a line network topology that forms a logical ring through sending back the frames by the last node due to the full-duplex links. Moreover, the star topology can be handled by using specific equipment, i.e. VABS ...
Context 24
... (Very High Performance Automation Bus System) is an academic RTE solution proposed in 2013 by [56]. Unlike many RTE protocol using the standard Ethernet frame to convey their data, VABS has its own frame format and its own data link layer protocol, as shown in Fig 2.12. This choice limits VABS compatibility with standard Ethernet devices. Figure 2.12: VABS frame structure VABS is based on the master/slave mechanism to convey synchronous traffic and a token passing for asynchronous traffic. As EtherCAT and SERCOS III, VABS supports a line network topology that forms a logical ring through sending back the frames by the last node due to the full-duplex links. Moreover, the star topology can be handled by using specific equipment, i.e. VABS ...
Context 25
... model the Profinet IRT cycle time, we consider the slipstreamimg effect [62,58,57,61], which consists in sending frames in the same physical order of the destinations from the master. Fig. 2.10 shows a spatio-temporal diagram of Profinet. The slipstreaming effect is beneficial and considered only when it is positive, which means τ ≥ δ + l , i.e., α ≥ 0. Consequently, the Profinet IRT MCT ...
Context 26
... [32], an interesting classification of the main RTE solutions has been detailed based on the implementation level of each proposed solution, as represented in Fig. 4.1. Hence, a first class with an implementation on top of the transport/network layer has been identified which does not require a special hardware, e.g., P-NET [39], V-NET [40,41] and Modbus-RTPS [42,43,44]. These solutions are usually easier to implement and configure, but they lead at the same time to important latencies (about 10ms), which make them more effective for soft real-time applications. Then, a second category has been defined, which provides a realization on top of the MAC layer while keeping the IEEE802.3 compatibility, e.g., TCNET [45], Ethernet/IP [46,47] with Device Level Ring (DLR) [48,49] and PowerLink [50]. Finally, the third category of solutions modifies the standard implementation, by introducing a set of hardware modifications, to guarantee a better real-time behaviour. A different classification is introduced herein to distinguish the main RTE solutions from an avionics perspective, as shown in Fig. 2.7. Four classes of RTE solutions supporting ring topology have been ...
