Fidia S.p.a. · MIL:FDA

Fidia — Technical Presentation (March 2003)

13 slides from Fidia's technical presentation, Mar 11, 2003

  1. Slide 1: OCEAN: Usability Study of Available Real-Time Operating and Communication Systems

    OCEAN Open Controller Enabled by an Advanced real-time Network IST-2001-37394 Usability Study of Available Real-Time Operating and Communication Systems Dipl.-Ing. Andreas Kahmen Laboratory for Machine Tools and Production Engineering Aachen University of Technology 11 March 2003

  2. Slide 2: Contents of presentation

    Contents of presentation  Analysis of available Linux implementations  Comparison of Linux real-time extensions  Investigation of Linux networking and device support  Comparison of Linux with standard operating systems  Analysis of available CORBA implementations  Comparison of CORBA implement

  3. Slide 3: Comparison of Linux implementations

    Comparison of Linux implementations  Suitable Linux real-time extensions are RTLinux and RTAI  The comparison resulted that RTAI provides better real-time support than RTLinux  soft real-time in user space along with hard real-time in kernel space  excellent performance in terms of low jitter an

  4. Slide 4: Linux Networking and Device Support

    Linux Networking and Device Support  Linux support of network devices  Ethernet (10 Mbps, Fast and Gigabit Ethernet)  FDDI (Fiber Distributed Data Interface, 100 Mbps)  Wireless LAN (Local Area Network)  ISDN (Integrated Services Digital Network)  WAN (Wide Area Network)  Linux networking pro

  5. Slide 5: Benefits of Linux

    Benefits of Linux  Linux has been ported to various platforms  High level development and supporting tools are available  Full control over source code and operating system functionality  Open to all standards  No royalties  High reliability 5

  6. Slide 6: Comparison of Linux with standard OS

    Comparison of Linux with standard OS  Interrupt latency time and jitter are comparable to standards real-time OS  The coupling of real-time part and non-real-time part is realised like for commercial real-time extensions  More real-time drivers are available open source than for commercial soluti

  7. Slide 7: Comparison of CORBA implementations (I)

    Comparison of CORBA implementations (I)  Research for 11 different ORBs applying the following criteria:  Open source availablity  CORBA version implemented  Real-time capability and implemented RT CORBA features  Available language bindings  Implemented CORBA services  Supported multithreadi

  8. Slide 8: Comparison of CORBA implementations (II)

    Comparison of CORBA implementations (II)  All ORBs except for JacORB and VisiBroker RT have been ported and tested on Linux platforms  All ORBs except JacORB offer Language C++ bindings  TAO and ROFES provide real-time capability  Real-time versions of ORBExpress, e*ORB and VisiBroker are availa

  9. Slide 9: Interoperability between different ORBs

    Interoperability between different ORBs  ORB interoperability is covered by OMG specifications since CORBA 2.0  Interoperability is secured by General Inter-ORB Protocol (GIOP)  Interoperable Object Reference (IOR) specifies standardised object reference format  All ORBs of the research are impl

  10. Slide 10: Investigation of CORBA Component Model

    Investigation of CORBA Component Model  CCM defines features and services for implementation, management, configuration and deployment of components with commonly used CORBA services  Application of CCM eases software reuse  CCM provides Quality of Service (QoS) capabilities  monitoring and cont

  11. Slide 11: Time measurements for TAO

    Time measurements for TAO  Test environment:  Hardware: Motorola PowerPC (MPC604) 333MHz  Operating System: VxWorks 5.4.2 Tornado 2.0.2  Compiler: GCC Version cygnus-2.7.2  ORB: TAO 1.2a release of OCI  Test conditions:  One server and one client  methods provided by the server:  call_doubl

  12. Slide 12: Measurements for "call_long_void()"

    Measurements for “call_long_void()” roundtrip times for call of method "call_long_void()" 70 60 50 40 time in µsec 30 20 10 0 0 100 200 300 400 500 600 700 800 900 1000 measurement no. Server priority: 6 Client priority: 7 Measurements for round trip times 12

  13. Slide 13: Conclusions for time measurements

    Conclusions for time measurements call_double_ call_double call_long_ call_double_ Method call void() _ long() void() sequence() arithm. mean 24,095 24,089 24,081 24,185 value in µsec  The round trip measurements reflect the minimum time for method calls using TAO  Additional measurements with dif

Fidia — Technical Presentation (March 2003) on 6ix