210 likes | 575 Views
Programmable Logic Devices in NASA. Survey Results. Kalynnda Berens, SAIC Kalynnda.Berens@grc.nasa.gov. Programmable Logic Included. Programmable Logic Controllers (PLC) Programmable Logic Devices Field Programmable Gate Array (FPGA) Application Specific Integrated Circuit (ASIC)
E N D
Programmable Logic Devicesin NASA Survey Results Kalynnda Berens, SAIC Kalynnda.Berens@grc.nasa.gov MAPLD 2002
Programmable Logic Included • Programmable Logic Controllers (PLC) • Programmable Logic Devices • Field Programmable Gate Array (FPGA) • Application Specific Integrated Circuit (ASIC) • System-on-chip (SOC) • Complex PLD (CPLD) • others MAPLD 2002
Concerns • PLCs are “programmed”, usually by engineers. Programming languages may be graphical. • PL device complexity is similar to software, with many of the same problems • PL devices are ultimately “hardware”, but often programmed like software • Current assurance activities may not be adequate for the complexity of these devices. MAPLD 2002
Surveys • Usage and Assurance surveys sent to engineers, assurance personnel, and others at NASA Centers • Distributing surveys was more difficult than anticipated • Targeted email requests produced adequate responses, and apparently some heated exchanges! MAPLD 2002
Usage Survey • Identification • Programmable Logic (PL) Information • Types of PL Used and what for • Respondent PL familiarity • PL responsibility • Testing • Assurance • Standards, Procedures • Configuration Management • Assurance organizations • Safety MAPLD 2002
Assurance Survey • Identification • Project Programmable Logic Assurance • Center Programmable Logic Assurance • Review • Testing • Auditing • Witnessing • Respondent PL familiarity MAPLD 2002
Usage Survey Responses 2 surveys from non-NASA projects MAPLD 2002
Who filled out the Surveys? • Using PL: • Electronic/Computer engineers – 15 • Other Engineers – 27 • Management – 9 • Software engineer – 1 • Other – 3 • Assuring PL: • Safety – 6 • Software Assurance – 1 • Other assurance - 2 MAPLD 2002
PL Used within NASA • PLC – 27 • FPGA – 26 • ASIC – 12 • EPLD, EEPLD – 24 • SOC – 6 • PAL - 5 • Others - 11 MAPLD 2002
Where is PL Used? • Facilities – 19 • Flight Systems – 21 • Ground Systems – 11 • Research – 7 • Facilities are nearly all PLC. • PLC used in 3 experiments and 2 ground systems MAPLD 2002
Respondent PL Duties • Use, program, and assure – 23 • Use and program – 10 • Do not use – 7 • Use, program, or assure – 2 each • Program and assure – 2 • Design – 3 • Other - 4 MAPLD 2002
PL Author • Hardware engineer – 31 • Software Engineer – 12 • Specialist – 9 • Center computer science – 2 • Outside source/contractor – 8 • Other – 10 MAPLD 2002
PL Programming MAPLD 2002
Standards None – 35 Project/Center – 12 National/International – 4 Development/Design Procedures None – 31 Project/Center – 18 National/International – 1 Project Guidelines Yes – 20 No – 33 Configuration Management Yes – 39 No – 15 Standards, Procedures, Guidelines, Oh my! MAPLD 2002
Who Tests? Engineer or HW team – 31 Team with Software – 8 Team (sw?) – 5 User – 1 Technicians – 3 Engineer w/ QA – 2 Other - 3 How Tested? Successful test – 26 Simulation – 16 Extensive testing – 16 Incremental testing - 13 Testing MAPLD 2002
Assurance Activities Other: Project reviews, Configuration management, FMEA MAPLD 2002
Safety MAPLD 2002
Interesting Comments • It is a serious mistake to equate VHDL programming to software. At best it is firmware, but for the most part there isn't a good name for programming FPGA logic. In a high performance design like our, the minute you forget that you are designing hardware and think you are writing software you fail. • Historically, existing standards have been outdated and limit the capability of firmware development. MAPLD 2002
Interesting Comments • It sounds like this survey is being taken to try and promote additional overhead and oversight. The core problems that NASA is experiencing stem from excessive overhead and out of control bureaucracy. Attempting to set up a review process for firmware would result in having a committee of less than knowledgeable people assuming technical authority over personnel who would in general be more capable than those reviewing the firmware. This would be a step in the wrong direction. MAPLD 2002
Interesting Comments • Small projects can't afford QA watching over designer's shoulder to see that he is doing his job properly. The use of Peer Reviews and Design Reviews should be adequate. Using Test Procedures and providing enough testing time would be of greater value of resources and added confidence. MAPLD 2002
Open Questions • How does NASA compare to other groups? • Is the level of assurance adequate for how the PL is being used within NASA? • Is process assurance required, as well as product assurance? • Is there a complexity threshold? • What are the standard practices in industry, the military, and other government agencies for PL assurance? MAPLD 2002