Upload
truonghanh
View
217
Download
0
Embed Size (px)
Citation preview
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 1 of 33 v 0.2 04/02/00
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COINCOMO Software Cost and Schedule Estimation Models
COCOTS Integration Challenges
A. Winsor Brown Ed Colbert
[email protected] [email protected]
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 2 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
Goals Present the new, extended, combined COCOMO Model: COINCOMO (Incremental Development) • COCOTS for COTS software with applications
• COSECMO for software systems with security
Show how RUP/MBASE and COINCOMO fit with the Incremental Commitment Model for systems development
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 3 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
Outline
COCOMO + COPSEMO (COINCOMO Base) Models
COTS and (some) Open Source Models
COSECMO with COCOTS and COINCOMO = the New [or Extended] COINCOMO
Incremental Commitment Models (ICMs)
ICM for Software with Extended COINCOMO
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 4 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
RUP/MBASE Application Development Model1 IRR
LCA
IOC
PRR
LCO
CCD
1 (efforts not to scale)
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 5 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
Building on the COCOMO II (CII) Base COPSEMO (COngruent Phase Schedule and Effort MOdel) [nee COnstructive Phase Schedule and Effort MOdel]
EP% = XP% of ECIIbase SP% = YP% of SCIIbase
• Extrapolation of Effort (E) and Schedule (S) from CII's − EI% & SI% of CII to Inception E and S − ET% & ST% of CII to Transition E and S
• Interpolation of Effort (E) and Schedule (S) from CII's − EE% & SE% of CII to Elaboration E and S − 100-EE% & 100-SE% (the rest) of CII's to Construction
NOTE: Percentages currently all based on experience Recalculate Persons needed per phase: P=PM/M
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 6 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
MBASE/RUP Concurrent Activities IRR
LCA
IOC
PRR
LCO
CCD
COCOMO II Estimaton Covers
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 7 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COCOMO II with COPSEMO (for IECT)
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 8 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COPSEMO: Phased Schedule and Effort Dist.
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 9 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COCOMO Models Represented in UML COCOMO.II.2003 (the software package) COINCOMO.2006 version of COCOMO.II.2003 • Has concept of stored alternatives for − Components (AKA Project) [shown] − EAF and SF Driver sets [not shown] − Sub-Components (AKA Modules) [not shown]
• Has four SLOC sources: − SLOC or FP converted to SLOC: New & New Open Source − Adaptation Only: New Reused & Previous Build Reused
COINCOMO version of COCOMO.II.2003 with COPSEMO
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 10 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COCOMO.II.2003 in UML
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 11 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COINCOMO's COCOMO.II.2003 in UML
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 12 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COINCOMO's COCOMO + COPSEMO in UML
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 13 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COTS and Open Source Software Today's Realities: • Many, many systems have COTS components
• Many systems use Open Source components which are treated like COTS (AKA pseudo-COTS): − Might put wrappers around it − Don't look at internals − Let the Open Source developers control its evolution
COCOTS calculates effort and schedule for systems with COTS (or pseudo-COTS) • Activity based estimates for Assessment and Tailoring
• Special "Glue Code" model for COTS to rest of system
• NOTE: COTS volatility effects not calculated!
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 14 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COCOTS: COTS with Assessment, Tailoring and Glue Code
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 15 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COCOTS + COCOMO + COPSEMO Issues: • How to determine relative anchor point dates? − Apply maximum calculated LCA date
(adjust COPSEMO S&E percentages to fit) − Reality is that they are probably politically based ο But don't do LCO or LCA too early; ο If "scheduled" IOC before calculated IOC, adjust Sched!
• How to allocate Assessment to phases? − Initial Assessment (or screening) to Inception? − Detailed Assessment and selection in Elaboration?
• How to allocate Tailoring to phases? − All to Elaboration: COTS selection required(?) for LCA? − Need another percentage factor, derived from experience,
to allocate parts to Elaboration and Construction.
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 16 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COSECMO Extension to COCOMO COSECMO, the COCOMO SECurity MOdel, Focus: Cost & Schedule Estimates (C&SE) • COCOMO covers Elaboration (E) & Construction (C) • COSECMO covers the increased costs and schedule
for security, spread over ο Inception (I) ο Elaboration (E) ο Construction (C) ο Transition (T)
COSECMO to help government and industry, using modern practices (e.g., the Common Criteria), predict cost of developing or acquiring secure systems. COSECMO model still evolving, implementation based on COINCOMO available to Affiliates since the fall of 2005
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 17 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COSECMO
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 18 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COSECMO Impacts on Schedule COPSEMO ranges of percentages2
COSECMO increases in schedule depend on EAL – suggestions for initial selections (no experience) • EALs 1 and 2: use normal
• EALs 3 and 4: use high end of ranges
• EALs 5, 6 and 7: beyond the ranges shown above 2 from Table A.5 of
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 19 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COSECMO increases in schedule for security SI% of CII with COSECMO (for Inception) • EAL 3 and 4: SI% = 25. Higher because of Security Risk
Analysis, Security Target Definition and Protection Profile SE% of CII with COSECMO (for Elaboration) • EAL 3 and 4: SE% = 28+. Higher because of need for
architecture completeness with evaluation of security. SC% of CII with COSECMO (for Construction) • EAL 3 and 4: 100-SE% = 72- (72 or lower) ST% of CII with COSECMO (for Transition) • EAL 3 and 4: ST% = 20+. Higher because of certification
time (on top of increase related to higher cost).
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 20 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COSECMO with COPSEMO distributions
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 21 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COINCOMO=COSECMO+COPSEMO+COCOTS
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 22 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
Incremental Commitment Models The Incremental Commitment Models (ICMs) for Life Cycle Processes
• For Systems (Human, Hardware and Software)
• ICM for Software Intensive Systems (ubiquitous hardware)
• ICM for Software (only) Systems
ICMs solve Spiral Model problems • Use spiral principles vs. diagram
• Relate to stakeholder commitments and values
• Make concurrency explicit
• Use risk to explicitly show go-backs and skips
• Provide view for handling mini-spirals
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 23 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
ICM LC Processes For Systems [of Systems]
DoD, General/DoD Milestones
ExplorationCommitment
Review
ECR
ValuationCommitment
Review
VCR/CD
ACR/A DCR/BOCR1/C1DCR2/B2
OCR2/C2DCR3/B3
ArchitectureCommitment
Review
DevelopmentCommitment
Review
OperationsCommitment
Review1
OperationsCommitment
Review2
Phases (EVADO)
Activities
Exploration Valuation ArchitectingDevelopment1Architecting2
Operations1Development2Architecting3
Operations2Development3Architecting4
Concurrent Risk-and-Opportunity-Driven Growth of System Understanding and
Definition
Evaluation of Evidence of Feasibility to
Proceed
Stakeholder Review and Commitment
Initial ScopingConcept
Definition, Investment
Analysis
System Architecting
Increment 1 Development
Increment 2 Architecting Rebaseline
Increment 1 Operations
Increment 2 Development
Increment 3 Architecting Rebaseline
Feasibility Rationales ... ... ...
High, butAddressable
Too High,Unaddressable
Negligible Risk?
High, butAddressable
Too High,Unaddressable
Negligible Risk?
High, butAddressable
Too High,Unaddressable
Negligible
Adjust Scope, Priorities, or Discontinue
Risk?
High, butAddressable
Too High,Unaddressable
Negligible Risk?
High, butAddressable
Too High,Unaddressable
NegligibleRisk?
Acceptable Acceptable Acceptable Acceptable Acceptable
...
...
...
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 24 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
ICM Showing Software in Systems DoD, General/DoD
Milestones
ExplorationCommitment
Review
ECR
ValuationCommitment
Review
VCR/CD
ACR/A DCR/BOCR1/C1DCR2/B2
OCR2/C2DCR3/B3
ArchitectureCommitment
Review
DevelopmentCommitment
Review
OperationsCommitment
Review1
OperationsCommitment
Review2
Phases (EVADO)
Activities
Exploration Valuation ArchitectingDevelopment1Architecting2
Operations1Development2Architecting3
Operations2Development3Architecting4
Concurrent Risk-and-Opportunity-Driven Growth of System Understanding and
Definition
Evaluation of Evidence of Feasibility to
Proceed
Stakeholder Review and Commitment
Initial ScopingConcept
Definition, Investment
Analysis
System Architecting
Increment 1 Development
Increment 2 Architecting Rebaseline
Increment 1 Operations
Increment 2 Development
Increment 3 Architecting Rebaseline
Feasibility Rationales ... ... ...
High, butAddressable
Too High,Unaddressable
Negligible Risk?
High, butAddressable
Too High,Unaddressable
Negligible Risk?
High, butAddressable
Too High,Unaddressable
Negligible
Adjust Scope, Priorities, or Discontinue
Risk?
High, butAddressable
Too High,Unaddressable
Negligible Risk?
High, butAddressable
Too High,Unaddressable
NegligibleRisk?
Acceptable Acceptable Acceptable Acceptable Acceptable
...
...
...
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 25 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
Overlaps Across Software Builds
Inception Elaboration Construction Transition
Inception Elaboration Construction Transition
Evolve During Transition [After Sw IOC]
Evolve After Architecture Complete
Incept. Elaboration Construction Transition
Inception Elaboration with Evol. Req. Construction Transition
I. Elab. Construction Transition
. . .
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 26 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
ICM for Software
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 27 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
Multi-Build COCOMO II COINCOMO Sums Across Builds
Build x
NewBuild x
ModifyBuild x
ModifyBuild x+1
Carried
NewBuild x+1
Carried
Build x+1 Build x+2
Carried
NewBuild x+2
etc.New,
Reused andCOTS
New,Reused and
COTS New,Reused and
COTSBox size notional for effort.
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 28 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COINCOMO = Multi-Build COCOMO II Guidance about how to "carry" forward
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 29 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
COINCOMO = Multi-Build COCOMO II Guidance about how to "carry" forward (cont.)
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 30 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
Background WinWin Spiral Model (WWSM) WWSM: CSCI577 Unrolled with Repeated Cycles
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 31 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
Background and Definitions (cont.) WinWin Spiral Model • Risk Driven Selection, Execution and Validitaion of
Activities and Products
• Feasibility "demonstration" needed to proceed
• Stakeholder concurrence to proceed at major milestones
• Life Cycle Process(s) Model Generator: Select and document/plan for next "rounds" a specific "Development Process Model".
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 32 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
Stylized WinWin Spiral with Activities Mapped to Original Spiral[radial dimension (cost) not to scale]
2a. EvaluateAlternativeswith respect toOC&Ps
1b. StakeholdersIdentify SystemObjectives, Constrains,& Priorities (OC&Ps)Alternatives SolutionsElements
1a. IdentifySuccess-CriticalStakeholders
2b.Assess,AddressRisks
3. ElaborateProduct andProcessDefinition4. Verify and Validate
Product and ProcessDefinitions
Stakeholders’
Commitment4
5
6
8
2
1
Stakeholders’Review
7
3
L COL CACC DI OC
Progress Through Steps
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com
COCOTS Integration Challenges
© 2007 A W Brown BES/MSEE & USC CSE MOD-4 Brown Paper COINCOMO.doc – 33 of 33 v1.0 - 05/08/07
USC
C S E University of Southern CaliforniaCenter for Software Engineering
WWSM: CSCI577 Unrolled With Repeated Cycles
EOCDInc. Cycle
alreadydoneI & E
Key: IRR - Inception Readiness Review LCO - Life Cycle Objective
LCA - Life Cycle Architecture RLCA- Rebaselined LCA
LCA
RLCA
CCD
Time Line
SemesterBreak
StartOf Fall
Semester
SpringSemester
Ends
Inc. Elaboration Construction Transition
LCO
IRR
RRR2
IOC
TRR1
T RR RR R2 1
EOCD
I E C TFall Elab. Cycle Const. Cycle Prelim. Trans.
SummerSemester
Ends
CCD- Core Capability Demo. IOC- Initial Operational CapabilityTRR- Transition Readiness ReviewPRR - Product Release Review
Enhancement Cycle?
Presented at the 2007 ISPA/SCEA Joint Annual International Conference and Workshop - www.iceaaonline.com