25th International Forum on COCOMO and Systems Software Cost Modeling Systems Engineering Desk Reference An Aid for Cost Estimators 2-5 November 2010 University of Southern California Los Angeles CA Sherry Stukes Henry Apgar Jet Propulsion Laboratory/ California Institute of Technology Integrated Ground Data Systems Ground Software Systems Engineering 4800 Oak Grove Drive MS 301-225 Pasadena CA 818.393.7517 (o) 805.402.8664 (c) sherry.a.stukes@jpl.nasa.gov President MCR Technologies LLC 390 N Sepulveda Blvd Suite 1050 El Segundo CA 424.218.1616 (o) 805.402.4232 (c) hapgar@mcri.com Copyright 2010. All rights reserved. Background SSCAG* Systems Subgroup Desk Reference product Represents industry “Best Practices” Useful to Cost Estimators Organized into six sections Contributors NASA (JPL, LaRC, MSFC) SAIC MCR Technologies USAF (SMC, AFCAA) USC Raytheon Design for Value Model Vendors (Galorath, PRICE Systems) *SSCAG (Space Systems Cost Analysis Group) is an International working group comprised of member organizations that develop estimating products for the space industry. SSCAG currently has four active subgroups: Hardware, Software, Risk, and Systems, supported by members from industry, Government, and the academic community. 2 Desk Reference Overview System Engineering Desk Reference Content Document Overview How to Predict and Evaluate Systems Engineering and SoS Costs Definitions Rules of Thumb “Top 10”Tools Lessons Learned and Estimating Examples Sources RedStar Library Constellation Program Model vendor research University research Contractor organizations 3 Defining our Terms System of Systems (SoS) is a collection of task-oriented or dedicated systems that pool their resources and capabilities together to obtain a new, more complex, 'meta-system' which offers more functionality and performance than simply the sum of the constituent systems.” SOS Systems Subsystems Assemblies System Engineering can be considered to include the pure engineering efforts to ensure that a number of subsidiary elements function together properly but also project/program management, integration and test, missions assurance and ground support elements sometimes called “SEPM” or Systems Engineering and Project/Program. The definition can be adapted to lower levels, including subsystems and assemblies. 4 SoS Hierarchy The hierarchal aspect of SoS is reflected in the fact that depending on how one defines system, almost any integration activity can be tagged as SoS An example of SoS hierarchy is the crew launch system for the NASA Constellation Program. System of Systems” consists of the Ares I launch vehicle system and the Orion crew capsule system These systems themselves are comprised of multiple systems * Reference – materials submitted by Andy Prince, NASA Marshall Space Flight Center. 5 Rules of Thumb Model Development Requirements are king. More parameters increase the explanatory power of the model, but too many parameters make the model too complex to use and difficult to calibrate. Not all requirements are created equal. Reuse is not free. Operational Scenarios may come first, but requirements will Break the problem and analysis into phases ultimately describe the system. over time; the right amount of granularity is important. Don't double dip. Let available data drive the application boundaries of the model. Nominal is the norm. Design the rating scale according to the phenomenon being modeled. If you're estimating a large project, personnel capability is Nominal. Some system characteristics are more likely to be cost penalties than cost savings. Most of your off-Nominal cost drivers should match your last Find your sea level. project. If you're going to sin, sin consistently. Model Calibration All calibrations are local. Calibrations fix chronic errors in over- or underestimation. Be skeptical of data that you did not collect. For every parameter in the model, 5 data points are required for the calibration. Don’t do more analysis than the data is worth. You need less data than you think, you have more data than you think. Use a combination of models to estimate total system cost. Avoid overlap between models. Estimate using multiple methods (analogy, parametric, etc.). Estimation Estimate early and often. Experts all disagree forever. Bound the options they are given to evaluate. People are generally optimistic. Model Usage A model is not reality. All models are wrong, but some of them are useful. Begin with the end in mind. 6 “Top 10” Tools Version SE Defined LCC Phase Levels WBS Defined SE Est. Method Price ES System Engineering Process A-F Subsystem, System User Defined Imbedded Algorithm Academic COSYSMO algorithm Price True Planner System Engineering A-D SoS EIA/ANSI 632 WBS SEER -H System Engineering and Integration A-F Subsystem, System User Defined Imbedded Algorithm System Engineering & Integration A-E System Yes - Configurable CER by mission type System Engineering (w/PM) C,D System Yes – Defined CER USCM Systems Engineering (w/in Program Level Cost) A-D System Yes – Configurable CER by Spacecraft type NICM (both models) Systems Engineering (estimated but not def.) B,C,D thru L+30 System Yes - Defined CER in two forms COCOMO Software Engineering in the System Context (Waterfall WBS) Waterfall and Mbase System Software Waterfall WBS and MbaseWBS COCOMO algorithm COSYSMO for SE Systems Engineering Effort A-D System, Systems Eng EIA/ANSI 632 WBS academic COSYSMO algorithm COSYSMO for SoSE Systems of Systems Engineering Effort C-E SoS DoD SEGuide for SoS academic COSYSMO algorithm 7 NAFCOM SSCM07 Cost Factors - NASA SOS Cost Modeling System-of-Systems Cost = 0.07411*(DDTE$)0.9993*(1.09585)M/LV where DDTE$ = Total DDTE* Cost, in 2006$, M/LV = Manned or Launch Vehicle (Yes = 1, No = 0) System-of-Systems Actual Vs. Estimated Good Quality Metrics = 93.1% SPE = 33.2% $3,000 Estimated (2006 $M) R2 $3,500 $2,500 $2,000 $1,500 $1,000 $500 $0 $0 $500 $1,000 $1,500 $2,000 $2,500 $3,000 $3,500 Actual (2006 $M) *DDTE – Design Development Test and Evaluation 8 Apollo Atlas II Brilliant Pebbles HST ISS Peacekeeper Pioneer Venus Saturn IB Saturn V Shuttle Skylab Titan IV Systems of Systems Cost $1,654.1 $234.6 $77.8 $113.3 $3,646.4 $265.6 $24.9 $325.1 $603.9 $597.8 $334.1 $516.5 Total DDT&E Cost % of DDTE $18,449.0 8.97% $3,040.4 7.71% $1,100.2 7.07% $1,592.6 7.12% $36,027.4 10.12% $10,006.7 2.65% $306.3 8.13% $3,776.7 8.61% $10,333.9 5.84% $18,152.6 3.29% $4,127.7 8.09% $5,178.0 9.97% 2 WBS Description / Notes System of Systems Level Systems Engineering, Integration and Test and Top Level Program Management of Architecture Space Segment Assets in orbit or interplanetary 2.1.3 2.1.3.2 Crew Expolration Vehicle (CEV) Command Module (CM) CM PMP 2.1.4 2.1.4.1 Orion Integration, Assy., & C/O Orion SE&I Orion PM Orion STO Orion GSE Service Module (SM) SM SE IT PM SPACE 2.1 2.1.4.2 SM PMP 2.2 2.2.1 2.2.2 International Space Station ISS Communications TDRS Space Segment 3 3.1 3.1.1 3.1.2 Launch Segment CLV First Stage Upper Stage 3.1.2.2 3.1.2.1 4 4.1 4.1.1 4.1.2 4.1.3 4.1.4 4.2 * Reference – data collected and analyzed by Dr. Christian Smart, SAIC, under contract to NASA Marshall Space Flight Center. WBS Title SOS 1 LAUNCH Factor-based approach - derived from real data (c. Smart, SAIC) WBS Element 4.3 MISSION/GROUND OPS Estimate Example – NASA WBS Number Upper Stage Avioncis Software Upper Stage Avionics Hardware Upper Stage Integration, Assy., and C/O Upper Stage SE&I Upper Stage PM Upper Stage STO Upper Stage GSE Mission/Ground Ops Ground Stations Launch Head Wallops Island, VA New Boston, NH Argentia, Newfoundland Deep Space Network (DSN) Stations Range Safety CM Prime Mission Product (HW + SW) Systems Engineering, Integration and Test and Top Level Program Management of SM SM Prime Mission Product (HW + SW) additional satellite(s) to support CxP ISS mission SRM Improvements for communicationg with ground and TDRSS. Comms to/from AF ROCC Provides -S-band and UHF air-to- ground MILA/PDL = Merritt Island & Ponce de Leon; Merritt has 14 antennae; Ponce de Leon has 3 antennae; both are dedicated to shuttle/CxP missions. 5m and 8m S-Band tracking; not dedicated site Canberra, Goldstone, Madrid Backuo provided by DoD ground-based radars 9 Recommendations Carefully consider the application Consider how SoS differs from traditional engineering systems and how this affects the estimator Supporting platforms are operationally, geographic, and managerial independent, as well as network-centric New acquisition concepts means we need new CERs, factors, and cost drivers Immaturity of concept means little cost data is currently available SOS cost drivers are unique and require considerations beyond traditional systems estimating Review available research and papers Current research by USC, DAU, SEI, NASA, Cranfield University (UK) Available papers from USC, MIT, IEEE, INCOSE 10 Lessons Learned Inconsistent parameter definitions Between models (need to adopt INCOSE standards and develop a mapping scheme) Between historic data (collected before standards were established) and new CERs Inconsistent WBS applications Are we double-counting the SOS costs by applying factors at each platform level? Inconsistent platform applications Do we use the same factors for hardware systems as well as software systems? Are space platform factors different from air and ground platform factors? Are mannedspace platform factors different from unmanned-space platform factors? Might be more useful if, in the near term, we rely on databases (and factors) rather than on statistical CERs 11 Publication Schedule Highly motivated, volunteer workforce Hosted in a collaborative work area Periodic “tag-up” telecons Current Schedule Review 1 – 31 October 2010 Review 2 – 20 November 2010 Materials to Editor – 30 November 2010 Complete DR first draft – 31 December 2010 Seeking volunteer reviewers Published Desk Reference will be available in April 2011! 12