SAE AS 1393-2010 Serial Hi-Rel Ring Network for Aerospace Applications (RingNet)《航空航天设施用连续Hi-Rel环形网络(RingNet)》.pdf

《SAE AS 1393-2010 Serial Hi-Rel Ring Network for Aerospace Applications (RingNet)《航空航天设施用连续Hi-Rel环形网络(RingNet)》.pdf》由会员分享，可在线阅读，更多相关《SAE AS 1393-2010 Serial Hi-Rel Ring Network for Aerospace Applications (RingNet)《航空航天设施用连续Hi-Rel环形网络(RingNet)》.pdf（123页珍藏版）》请在麦多课文档分享上搜索。

1、AEROSPACE STANDARD AS1393 Issued 2010-03 Reaffirmed 2014-10 Serial Hi-Rel Ring Network for Aerospace Applications (RingNet)RATIONALE AS1393 has been reaffirmed to comply with the SAE five-year review policy. _ SAE Technical Standards Board Rules provide that: “This report is published by SAE to adva

2、nce the state of technical and engineering sciences. The use of this report is entirely voluntary, and its applicability and suitability for any particular use, including any patent infringement arising therefrom, is the sole responsibility of the user.” SAE reviews each technical report at least ev

3、ery five years at which time it may be revised, reaffirmed, stabilized, or cancelled. SAE invites your written comments and suggestions. Copyright 2014 SAE International All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by

4、 any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE. TO PLACE A DOCUMENT ORDER: Tel: 877-606-7323 (inside USA and Canada) Tel: +1 724-776-4970 (outside USA) Fax: 724-776-0790 Email: CustomerServicesae.org SAE WEB ADDRESS: http:/www.s

5、ae.org SAE values your input. To provide feedback on this Technical Report, please visit http:/www.sae.org/technical/standards/AS1393 INTRODUCTION Application perspective The bandwidth and resolution of aerospace remote-sensing payloads continue to advance, placing ever increasing demands on onboard

6、 data-handling networks. At the same time, economics and rapid development requirements are driving onboard data-handling networks toward flexible, nonproprietary architectures and interface standardization. These characteristics are extensively addressed in existing ground-based net- works. However

7、, unlike most ground-based networks designed primarily to support the transfer of non-real- time data between computers, an onboard data-handling network must support the real time needs of the aerospace remote-sensing environment. Network perspective Aerospace remote-sensing data is characterized b

8、y synchronous components, common to continuous mode sensors, and asynchronous components, common to event driven sensors. Both sensor types have real time data-handling requirements, and sensor performance is driving data bandwidth requirements into multiple gigabit per second range. Even at these h

9、igher data rates, aerospace systems will continue to be constrained by size, weight, and power limitations. In order to achieve high data-rate performance while maintaining low size, weight, and power, onboard data-handling subsystems must employ highly integrated components and must avoid the highe

10、r layer protocol features commonly found in ground-based networks. Furthermore, space-based data-handling networks must be fault tolerant and able to withstand the rigors of launch and the harsh space environment. Adaptation of existing ground-based network standards is not practical because of insu

11、fficient bandwidth and high power dissipation. Also, conversion of current implementations to space qualifiable processes is not feasible. Therefore, a new onboard data-handling network standard for aerospace remote-sensing applications is necessary. Technology perspective Fiber Optic and ASIC techn

12、ologies have matured to the point that multiple gigabit data-handling networks for aerospace applications are practical. Fiber Optic technology is especially appealing because of its extremely high bandwidth capacity. Multimode fibers have been demonstrated to be radiation hard with negligible incre

13、ase in signal loss over a ten year life in space radiation environments. Space qualifiable connectors have been designed and built specifically for fiber cable. Techniques for low loss termination and coupling of fibers have been perfected. High power laser diodes capable of operation over a wide te

14、mperature range and low noise optical receivers are now available. State-of-the-art semiconductor materials and processes, as well as packaging techniques, are leading towards smaller, faster, and lower power devices that will enable high-speed data-handling systems for space applications. TABLE OF

15、CONTENTS 1. SCOPE . 9 1.1 Purpose 9 2. References . 9 2.1 Applicable Documents . 9 2.2 List of Acronyms . 9 2.3 Definitions 10 3. Network Description . 12 3.1 Overview 12 3.1.1 Topology 12 3.1.1.1 Addressing . 12 3.1.2 OSI Protocol Layer Allocation 12 3.1.3 Network Characteristics . 14 3.1.3.1 Physi

16、cal Layer Characteristics . 14 3.1.3.2 Data Link Layer Characteristics . 14 3.1.3.3 Management Layer Characteristics . 16 3.2 Operation . 17 3.2.1 Control Node Interface . 17 3.2.1.1 Discrete Command and Control Interfaces 17 3.2.1.2 Register Interfaces . 18 3.2.1.3 CCIR Structure and Definition 19

17、3.2.1.4 CIU Single Command Definitions 22 3.2.1.5 CIU Block Command Definitions 24 3.2.1.6 Block Load Command Execution . 26 3.2.1.7 Initialize Command i.e., the fiber optic cables and connectors. The BIU, or Bus Interface Unit, is intended to be an embedded module within each Data Node. It provides

18、 the primary data handling functions required to both send and receive data over the RingNet network. The CIU, or Control Interface Unit, is intended to be an embedded module within a Control Node. It provides the primary configuration, control and status monitoring interface to the RingNet network.

19、 FIGURE 3-1 - RINGNET TOPOLOGY 3.1.1.1 Addressing Address 0x7F is reserved for broadcast commands and address 0x7E is reserved for the optional Bus Monitor function, limiting the number of BIUs to 125. Address 0x00 is reserved for the CIU. The remainder of the addresses (0x01 0x7D) can be used for B

20、IU addresses. 3.1.2 OSI Protocol Layer Allocation In an effort to standardize the protocol structure, the International Standards Organization (ISO) created the 7layer Open System Interconnection (OSI) model to define and partition the basic network processes (see Table 1). The RingNet network incor

21、porates the three layers of this 7layer model associated with data transfer, the Physical Layer, Data Link layer and Network Layer. The RingNet also provides a limited set of Network Management functions. These Network SAE INTERNATIONAL AS1393 Page 12 of 123_ Management functions have been combined

22、with the Network Layer functions to form the RingNet Management Layer. Figure 3-2 illustrates in a simplified form the allocation of these protocol layers to the RingNet functional element. Network Independent Layers 7- Application Layer Management Layer 6 - Presentation Layer 5 - Session Layer 4 -

23、Transport Layer Network Dependent Layers 3 - Network Layer 2 - Data Link Layer 1 - Physical Layer 1.1 - Non-Media Dependent Physical Layer 1.2 - Media Dependent Physical Layer TABLE 1 - ISO/OSI MODEL a) Physical Layer - In the RingNet the Physical Layer is divided into two parts, the Media Dependent

24、 Physical Layer (MDPL) and the Non-Medial Dependent Physical Layer (NMDPL). The passive elements of the MDPL, the network cable and connectors, are contained within the Physical Plant. The active elements of the MDPL associated with signaling and the NMDPL processes associated with symbol encoding a

25、nd timing are contained within the BIU and CIU. b) Data Link Layer - The Data Link Layer (DLL) functions associated with data formatting, data transfer and the Data Node interface are provided by the BIU. The DLL functions associated with network synchronization and frame formatting are provided by

26、the CIU. No Network Independent Layers are provided by the RingNet. c) Management Layer - In the RingNet, the Network Layer functions associated with network bandwidth allocations and data routing, are considered part of the Management Layer. These non-real time Management Layer functions are provid

27、ed by the Control Node with the CIU serving as the command and status interface between the Control Node and the RingNet network. FIGURE 3-2 - PROTOCOL LAYER ALLOCATION SAE INTERNATIONAL AS1393 Page 13 of 123_ 3.1.3 Network Characteristics 3.1.3.1 Physical Layer Characteristics The baseline configur

28、ation for the RingNet network is a redundant, cross-strapped, serial ring with a passive optical bypass feature (see Figure 3-3). The network is capable of supporting up to 125 BIUs and a CIU with a maximum node-to-node spacing of 100 meters. Section 3.1.1.1 for explanation of the number of nodes. T

29、he Media Dependent Physical Layer is characterized by multimode, graded index fiber with laser diode transmitters and PIN diode receivers operating at an optical frequency in the 1300 nm range. Although multi-fiber cable, multi-pin connectors and specific termini are recommended for inter-operabilit

30、y, the user is free to select a different cable interconnect scheme. The Non-Media Dependent Physical Layer is characterized by a continuous transmission, serial communications mode. Data is encoded using 8B/10B symbol encoding and frame level synchronization is maintained using the unique K28.5 and

31、 K28.7 8B/10B command codes. The encoded symbols are transmitted serially using direct modulation digital On/Off signaling. The RingNet supports a selectable data rate of 100 Mbps to 2.5 Gbps. FIGURE 3-3 - RINGNET REDUNDANCY CONFIGURATION 3.1.3.2 Data Link Layer Characteristics The Data Link Layer s

32、ervices are provided by the protocol processing elements of the BIU and CIU. These services include the Data Node-to-BIU and BIU-to-Data Node data transfer services, the ATM Cell formatting services, the RingNet media access services and the CIU Latency Adjust Buffer services. Transfer of data betwe

33、en Data Nodes on the RingNet ring is accomplished using a simple fixed length 32-Slot, TDMA frame format. Each of the 32-Slots are a fixed 56 bytes in length and contain a standard 53-byte ATM Cell and 3 bytes of RingNet control overhead (see Figure 3-4). There is a deterministic delay around the ri

34、ng equal to one, two, three or four frames depending on the number of BIUs and the distance between the BIUs in a particular network configuration. This deterministic delay is maintained by a Latency Adjust Buffer (LAB) within the CIU that compensates for BIU quantity and spacing variations and insu

35、res that an integral number of Frames is rotating on the ring. Network bandwidth can be dedicated to a BIU using the Dedicated Transmit Slot process or shared between multiple BIUs using the Token Arbitrated process. Both processes transmit data in a broadcast mode. Data delivery can be connection o

36、riented using the Dedicated Receive Slot process or packet oriented using the ATM Header Address process. Further explanation of these processes is provided later in this subsection. SAE INTERNATIONAL AS1393 Page 14 of 123_ FIGURE 3-4 - RINGNET FRAME FORMAT The transfer of an ATM Cell between the BI

37、U and the Data Node forms the DLL boundary. The Data Node is expected to provide all message layer and higher layer OSI protocol services. In the special case of continuous unformatted data transfer between the BIU and its Data Node the BIU can perform the data segmentation and reassembly function n

38、ormally attributed to the Message Layer. The following items summarize the key characteristics of the RingNet Data Link Layer. Data format - The RingNet network is capable of transferring continuous, unformatted data and formatted data packets between BIUs. Both data types can be transferred synchro

39、nously or asynchronously. The format of the packetized data can be either simple 48-byte blocks or fully formatted ATM Cells. Data delivery - A RingNet network supports both circuit switched and packet switched data delivery services. These services may be mixed on the same RingNet network. Both ser

40、vices use multicast as the primary means of data transfer so that data destined for multiple BIUs is transmitted only once. Data throughput - The RingNet network is scalable to allow the network to be optimized to meet specific spacecraft data throughput requirements. The network elements are scalab

41、le to accommodate a maximum node-to-node data rate of 100 Mbps to 2.5 Gbps. Bandwidth reuse - In a structured spacecraft architecture, the data handling capacity of a RingNet network can be multiplied several times by utilizing the RingNets Bandwidth Reuse feature. Bandwidth Reuse simply means that

42、once data reaches the destination BIU the bandwidth used to transport that data can be immediately reused by the destination BIU or any succeeding BIU. Deterministic latency - Another key feature of a RingNet network is the deterministic nature of the data transfer latency. The data transfer latency

43、 on a RingNet network is deterministic and stable to within one bit. This is a product of the fixed-length frame format and the Latency Adjust Buffer (LAB) within the CIU. Data transfer options - The RingNet network offers the user several methods of transferring data between BIUs. In fact, the abil

44、ity to transfer both synchronous and asynchronous data, using either dedicated or token arbitrated bus bandwidth, in either a circuit switch or packet switch network configuration, can present a confusing set of options. However, all combinations of data transfer methods available on a RingNet netwo

45、rk can be simplified into two methods of sending data and two methods of receiving data. Sending data - Network bandwidth is allocated by assigning one or more of the 32-Slots of a RingNet Frame to each of the BIUs. These Slots can be either dedicated to a specific BIU or shared between multiple BIU

46、s. When Slots are dedicated, the BIU can insert data onto the network each time the assigned Slots rotate pass the BIU. When Slots are shared, the BIUs utilize token passing to arbitrate access to the assigned Slots with other BIUs. The Dedicated Transmit Slot process is primarily intended for the t

47、ransmission of continuous, high rate sensor data. The Token Arbitrated process allows multiple sources to efficiently share one or more Slots on the bus. Both transmit processes SAE INTERNATIONAL AS1393 Page 15 of 123_ are broadcast in nature. Slot assignment is dynamically managed by the Control No

48、de using the RingNet Subframe OH allowing real time bandwidth management without reducing data throughput. Receiving data - The broadcast data is accepted or rejected using either a Dedicated Receive Slot process or a ATM Header Addressed process. The Dedicated Receive Slot process supports the circuit switched method of data transfer. The ATM Header Addre