Pds performance and dependability symp 2004 ieee

This preview shows page 87 - 96 out of 114 pages.

PDS: Performance and Dependability Symp. 2004: IEEE Trans. Dependable and Secure Computing Deep submicron effects F-T ARM, Sparc Intrusion detection and defense Widespread use of server farms for directory sites and e-commerce (Google, e-Bay, Amazon)
Nov. 2019 Part VII – Failures: Computational Breaches Slide 88 Dependable Computing in the 2010s 2010: DSN Conf. held its 40th in series, going back to FTCS Cloud computing reliability focus of an IEEE TDSC 2013 special issue Greater integration of reliability and security concerns End of decade: Half-century mark of the field and of the DSN Conf.
Nov. 2019 Part VII – Failures: Computational Breaches Slide 89 A.2 Long-Life Systems Genesis: Computer systems for spacecraft on multiyear missions, with no possibility of repair Today: More of the same, plus remotely located, hard-to-access systems for intelligence gathering or environmental monitoring Typical systems for case studies: NASA OAO, Galileo, JPL STAR, . . . , Space Station Communication satellites Remote sensor networks
Nov. 2019 Part VII – Failures: Computational Breaches Slide 90 The JPL STAR Computer Became operational in 1969, following studies that began in 1961 Standby redundancy for most units, 3 + 2 hybrid redundancy for TARP mod-15 inverse residue code to check arithmetic ops and bus transfers Also used other codes, plus various hardware/software sanity checks Bus checker Test and repair processor
Nov. 2019 Part VII – Failures: Computational Breaches Slide 91 A.3 Safety-Critical Systems Genesis: Flight control, nuclear reactor safety, factory automation Today: More of the same, plus high-speed transportation, health monitoring, surgical robots Typical systems for case studies: CMU C.vmp Stanford SIFT MIT FTMP August Systems High-speed train controls Automotive computers
Nov. 2019 Part VII – Failures: Computational Breaches Slide 92 Avionic Fly-by-Wire Systems Source: [Trav04] Airbus A320 entered operation in 1988 Other models include A340 and A380 Primary (P) and secondary (S) computers (different designs and suppliers) Multiple redundant software modules Architecture of one computer, with its command and monitoring sections Three hydraulic systems
Nov. 2019 Part VII – Failures: Computational Breaches Slide 93 Automotive Drive-by-Wire Systems Source: [Seid09] Input Actuation Processing Fully electronic, with redundant buses, and power supplies Interim braking solution with mixed electronics and hydraulics
Nov. 2019 Part VII – Failures: Computational Breaches Slide 94 A.4 High-Availability Systems Genesis: Electronic switching systems for telephone companies Today: More of the same, plus banking, e-commerce, social networking, and other systems that can ill-afford even very short down times Typical systems for case studies: AT&T ESS 1-5, telephone switching, 1965-1982 Tandem NonStop I/II- . . . -Cyclone-CLX800, 1976-1991 Stratus FT200-XA2000, 1981-1990 Banking systems Portals and e-commerce (Google, Amazon)
Nov. 2019 Part VII – Failures: Computational Breaches Slide 95 Tandem NonStop Cyclone Announced in 1989 for database and transaction-processing applications (descendant of the first NonStop system announced in 1976) Mirrored disks, with multiple access paths Unidirectional optical links Redundant channels, I/O buses, controllers

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture