A low maintenance and highly reliable fly-by-light architecture

Fiber optic systems are often cited as a way to gain EMI immunity, reduce weight, and increase bandwidth. Avionics Full-Duplex Switched Ethernet (AFDX), specified in ARINC 664 Aircraft Data Network, Part 7, is a fiber optic avionics bus specification used on the Airbus A380, Boeing 787, and others. It was designed as an upgrade to the standard Ethernet protocol which added guaranteed determinism with bounded jitter and latency. This was done to allow hard real time, critical systems to communicate using standard IEEE 802.3 Ethernet protocol. It uses dual redundancy and full duplex links to minimize jitter and latency and eliminate packet collisions. Although AFDX was designed with hard real time systems in mind, it has not yet been used in a safety critical flight control system. Airbus has expressed interest in extending its use of AFDX from mission critical to flight critical systems[1]. The use of AFDX for flight control systems could provide a number of benefits over legacy systems such as MIL STD 1553 and ARINC 429 (A429). Other fiber optic busses do exist, including MIL STD 1773, ARINC 629, ARINC 636 and Fiber Channel, but these are not currently used for flight controls and comparisons to these busses is beyond the scope of this study.

We examine in this paper the reduction in receiver sensitivity in passive fibre optic star bus networks due to the laser bias power transmitted by every terminal. Both transimpedance and high impedance preamplifier configurations using PIN and avalanche photodiodes are considered. The penalty in receiver sensitivity is examined as a function of bit rate, number of terminals, amplifier noise and the ratio of the launched laser bias power to average launched signal power. For PIN receivers the penalty has a maximum at low bit rates and falls off rapidly on either side of the maximum. In an example quoted in the paper the PIN receiver sensitivity penalty for networks with less than 50 terminals is less than 1 dB for all bit rates. The penalty for APD receivers increases with bit rate, and, in a further example, is shown to be less than 1 dB for networks with less than 50 terminals operating at bit rates lower than approximately 20 Mbit s−1. Techniques for reducing the penalty to negligible levels are also examined.

We present the key features of OPTIMUL (OPTically Interconnected MULtiprocessor) — a multiprocessor architecture that implements a distributed shared memory mechanism using a wavelength division multiplexed (WDM) fiber optic bus. OPTIMUL also supports distributed memory style multiprocessing using the same high-speed message passing mechanism that implements the distributed shared memory. Trace-driven simulation results seem to indicate that efficient performance in the shared and distributed memory modes of operation stem from the operational concurrency of the WDM channels and the high data rate and low latency possible from the use of the new fiber optic receivers and transmitters.KeywordsShared MemoryWavelength Division MultiplexFiber LinkHome NodeCache BlockThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

The architecture and design of a Modular Distributed Avionics System (MDAS) for launch vehicles are presented which attempt to lower costs in several ways: use of commercial components and processes; common module design which includes flexible sensor signal conditioning and flexible valve control; reduction of cabling cost and weight; elimination of power conditioning requirements; reduction of battery requirements; use of ISO 11898 CAN serial command and data bus; total programmability via umbilical; standard software with all commands and functions that are customized to a specific timeline via an overlay of a high-level of the operational sequence and timeline.

This paper describes the operation of a new type of fiber optic local area network (LAN) that has most of the advantages of both the contention and token systems. The new system, called PRENET, is basically a contention system; it has no collisions, no message size or cable length limitations, and no stability problems. It can be loaded up to 80-85 percent of its bus capacity and it has bounded delays. The PRENET system has active stations and can be operated with two or four fiber optic cables. When operating with four cables the advantage of PRENET over other networks is its robustness against station failures and open-cable problems. The PRENET system uses a new media-access protocol called Preemptive Contention. It is shown that with this new protocol a packet- or station-based priority scheme can be readily included in the operation of the system. The computer simulation of a two cable system using the Preemptive Contention protocol is also presented. It is shown that for such a system operating at 2 Mbits/s with 80 percent bus utilization the mean waiting time is under 1 ms. It is also shown that the waiting time of a four-cable PRENET system is still smaller. It is pointed out that the new media-access protocol of the PRENET simplifies the behavior of this contention system to one that can be approximately predicted using an M/D/1 queueing model with batch arrivals.

E-Systems/ECI Division has developed a fiber optic based broadband communications bus for application to ground airborne and shipboard Command Control Communication Intelligence (C31) platforms. The system is based on a linear low noise fiber optic backbone bus with up to 3 GHz bandwidth and interface modules which allow voice (PABX) data and video to be handled and switched in one common media. This concept greatly reduces the complexity of system integration and deployment while decreasing the system weight.

We review the recent results on the design and implementation of the performance-optimized bidirectional optical backplane bus aimed at high speed multiprocessor systems. It employs an array of multiplexed holograms, in conjunction with a waveguiding plate within which cascaded fanouts are generated. Data transfer rate of 1.2 Gbit/sec at 1300 nm is demonstrated with a single bus line for a system composed of nine boards. Packaging-related issues and misalignment effects are addressed. Theoretical treatment to minimize fluctuations among the received powers at each board was carried out and the overall performance was optimized. We also introduce a hybrid optical backplane with multiple bus lines. The hybrid backplane maintains the same waveguiding structure, and exploits optoelectronic array devices. A new demonstration system is being implemented. The optical backplane that we developed is transparent to higher level bus protocols, thus can support standard backplane buses such as Futurebus+, Multibus II, and VMEbus.

Federated learning (FL) is a highly promising collaborative machine learning method that preserves privacy by enabling model training on client nodes (e.g., mobile phones, Internet-of-Things devices) without sharing raw data. However, FL is vulnerable to Byzantine nodes, which can disrupt model performance, render training ineffective, or even manipulate the model by transmitting harmful gradients. In this paper, we propose a Byzantine fault-tolerant FL algorithm called federated learning with trustworthy data and historical information (FLTH). It utilizes a small trusted training dataset at the parameter server to filter out gradient updates from suspicious client nodes during model training, which provides both Byzantine resilience and convergence guarantee. It further introduces a historical information-based credibility assessment scheme such that the client nodes performing poorly over the long-term have a lower impact on the aggregation of gradients, thereby enhancing fault tolerance capability. Additionally, FLTH does not compromise the training efficiency of FL because of its low time complexity. Extensive simulation results show that FLTH achieves higher model accuracy compared to state-of-the-art methods under typical kinds of attack.

This paper presents measured performance data for linear fiber optic data bus configurations. Requirements and measurements for the individual data bus components including fiber optic couplers, transmitters, and receivers are given, including environmental performance. Demonstrated performance parameters for several linear data bus configurations with up to 69 nodes and data rates up to 50 megabits per second are presented.

Since 1998 the first optical data buses, based on a ring topology and polymethylmethacrylate (PMMA) fibers with 1 mm core diameter and red emitting LEDs were introduced in cars to interconnect information and entertainment components. These systems currently are limited to the passenger compartment, because PMMA fibers do not tolerate environmental temperatures above 85 /spl deg/C, the data rate of these systems is limited to well below 100 MBit/s and the ring topology does not allow the interconnection of safety critical control units. Optical datacom buses based on standard multimode fibers are not useable in cars because of the severe environmental conditions and because of the extremely low-cost demands for in car data transmission systems. We propose and demonstrate a star coupled data bus system based on 200 /spl mu/m core diameter hard clad silica (HCS) fibers for 16 terminals with a data rate of up to 200 MBit/s.

The Navy's commitment to the application of data bus technology to shipboard internal communication stems from the necessity ofreducing cable congestion and weight. The Shipboard Data Multiplex System is a general purpose data bus system intended to take the placeof most of the point -to -point cabling and associated hardware presently used for information transfer aboard ships of the Fleet. It uses bothfrequency and time multiplexing of a two -level set of redundant interconnections to provide a highly reliable data bus capable of distributingall forms of periodic and aperiodic signals.We report here a demonstration of the feasibility of substituting fiber optics for the coaxial cable presently used by this system. Primaryconcerns were whether the interface hardware could handle the FDM signals on the bus without excessive distortion and degradation of theirsignal to noise ratios.This paper describes methods used to demonstrate that the intermodulation distortion and signal -to -noise ratio requirements of theSDMS can be met by the necessary interface components.IntroductionThis paper reports the first phase of an effort to demonstrate and evaluate the feasibility of applying fiber optics technology to the ad-vanced development model of the Shipboard Data Multiplex System (SDMS), a multiterminal multiplexing (data bus) system. SDMS is in-tended to serve as a general purpose information transfer system taking the place of most of the point -to -point cabling and associated hard-ware presently used for information transfer aboard naval ships.1 , 2 It has been built by the Autonetics Division of Rockwell Internationalunder contract to the Naval Sea Systems Command. This new system uses frequency and time multiplexing of a two -level set of redundantinterconnections to provide higher reliability and survivability with lower weight than present point -to -point systems. It is potentially capableof distributing throughout a ship all forms of periodic and aperiodic signals, either analog or digital.Two -level multiplexing, wherein user access is provided through remote multiplexers (RMs) which transfer information to area multi-plexers (AMs) and thence to a common data bus, allows many access points to be serviced without the attenuation that would result fromseparate connections to a common data bus. The first stage of multiplexing, consisting primarily of pulse coding in the RMs, is followed bytime multiplex transmission to the AMs at base band. Each AM serves up to eight RMs on a direct basis. Each dual redundant RM is

In this paper, we propose several modulation techniques for use in military avionics optical fiber data buses, namely, Extended Manchester II Bi-Phase Coding with Beginning-Stopping Flags, Partial Trilevel Manchester II Bi-Phase Coding, Four-ary Pulse Width Modulation, Two-ary Pulse Width Modulation with Amplitude-Distinguished Sync Field, Two-ary Message Pulse Width Modulation and Two-ary Sync Pulse Amplitude Modulation, respectively. Compared with an existing modulation scheme of MIL-STD-1773 avionics data buses, the proposed techniques can be effectively used to overcome the difficulty in recognizing the correct operation states of an active transmitter at the output of optical receivers. The feasibility of proposed modulation schemes are discussed, and their performance is also evaluated.

This paper demonstrates how to design a fault tolerant web application that relies on an affordable redundancy in data and function using Simple Object Access Protocol (SOAP). The proposed approach relies on a Single-Version fault tolerance model without using multiple versions or the N-Version fault tolerance model. The work aims to design a more dependable and reliable Web Service system at no extra cost on multiple versions.

The drastically reduced costs and improved performance of fiber optic components in recent years offers an opportunity to economically incorporate high-bandwidth optical communication links within a large-scale multiprocessor system. In this paper, the design, implementation, and initial performance of a fiber optic bus for a tightly coupled multiprocessor system are discussed. This optical bus ulitizes a multiple fiber-multiple star configuration with data rates of up to 80 MHz. It transmits bus control line information by repetitively scanning the control lines and transmitting their states as a packet of bits. It is currently being used for a 30-processor system called the Heidelberg Polyp, which has a custom bus design. However, the design approach is a general one that could easily be adapted to other buses, especially asynchronous buses such as the VME bus.

The digital era has witnessed an unprecedented surge in data generation, driven by diverse sources such as social media platforms, e-commerce, Internet of Things (IoT) devices, mobile applications, and enterprise systems. This explosive growth has led to the emergence of big data, characterized by the five Vs: volume, velocity, variety, veracity, and value. Managing such massive and complex datasets necessitates the use of scalable storage systems that can efficiently store, control, and retrieve data. Traditional storage infrastructures—such as centralized databases and file systems—struggle to meet the demands of big data applications due to limitations in scalability, fault tolerance, data redundancy, and performance bottlenecks. There are lot of problems with old system. In response, modern storage architectures have evolved to embrace distributed and cloud-based storage systems that offer horizontal scalability and high availability. The exponential increase in data generation across industries has intensified the demand for storage architectures that are both scalable and efficient. Conventional systems often face challenges in sustaining high ingestion rates, minimizing retrieval latency, and adapting to the dynamic requirements of large-scale data analytics. This research focuses on optimizing scalable storage architectures to enhance the performance of big data processing frameworks. It examines distributed file systems, object-based storage, and cloud-native approaches with particular attention to data distribution, replication mechanisms, metadata handling, and resource utilization. The study also analyzes the balance between scalability, fault tolerance, and consistency, while considering integration with platforms such as Hadoop and Spark. Through systematic benchmarking and performance evaluation, the proposed models aim to improve throughput, reliability, and cost-effectiveness. The outcomes of this research are intended to guide the development of next-generation storage solutions capable of supporting the continuous growth of big data ecosystems.