Technology Obsolescence Forecasting and
Management
Electronic Systems Cost
Modeling Laboratory
http://www.enme.umd.edu/ESCML/
Peter A. Sandborn
Design Refresh Planning
Introduction
• Obsolescence is defined as the loss or impending loss of
original manufacturers or suppliers of items or raw
materials (also known as DMSMS - Diminishing
Manufacturing Sources and Material Shortages).
• Technology obsolescence is a significant contributor to
the high sustainment costs of complex systems.
• As sustainment-dominated systems are forced to use
more COTS (Commercial Off the Shelf) technology,
obsolescence problems become more significant.
MOCA Results
• To perform design refresh planning,
a timeline of obsolescence,
mitigation, production, and design
refreshes for hardware and software
must be modeled:
Investment
20%
• Expansion
• Retrofit
• Spare replenishment
• Other planned production
Nothing is obsolete
Each data point
represents a
candidate design
refresh plan for the
system
Design refresh
Redesign nonrecurring costs
Functionality
Upgrades
Re-qualification?
Re-certification?
EMD
Parts to be replaced on
a particular board at a
particular refresh
Hardware and Software
• Lifetime buy
Sustainment
Production
2013
“Long term”
mitigation strategy
• Substitution
• Emulation
“Short term”
mitigation strategy
• Existing stock
• Last time buy
• Aftermarket source
2008
Sustainment
78%
2003
Part or technology
becomes obsolete
Example Life Cycle Curve for a 16M DRAM
2000
1500
1000
500
2000
93
94
95
96
97
98
99
00
01
02
03
04
Year
By curve fitting part sales data,
trend equations can be
formulated to forecast the
obsolescence dates of current
and future electronic parts.
(64M, May 2000)
0.0011
µ =1991.8M
(16M, March 1998)
MOCA Design Example
The Mitigation of Obsolescence Cost Analysis (MOCA) tool was
used to determine the optimum design refresh dates and refresh
content for an avionics module subject to electronic part
obsolescence.
1998
1996
(4M, February 1997)
1994
1992
(1M, March 1991)
1990
0
92
Life Cycle Trends for
DRAMS
2002
16M actual
16M forecast
05
1988
1986
0
(0.26M, May 1998)
10
20
30
Standard Deviation of Gaussian Fit to Sales Data, σ
Life cycle profile parameters:
µ = 1998.2
σ = 1.6 years
Peak Sales Year, µ
2500
Current Date (July 1999)
3000
Units shipped (in millions)
Most electronic part
manufactures are driven
by high-volume
commercial applications,
e.g., mobile phones.
Therefore, parts often
have lifetimes that are
far shorter than that of
sustainment-dominated
systems, e.g., avionics.
Part or technology is obsolete
short term mitigation strategy used
Start of Life
2018
F-22 Life Cycle
Cost (projected)
The F-22 program has spent over $80M on electronic part
obsolescence resolution already
Obsolescence
Forecasting
The MOCA software tool determines the optimum design
refresh date(s) based on:
• Forecasted technology obsolescence (what and when)
• How obsolescence events are mitigated
• Production, retrofit and sparing requirements
1998
Obsolescence affects:
• Electronic parts
• Materials
• Software
• Intellectual property
• Requirements and
specifications
R&D
2%
• Design refreshes are performed on sustainmentdominated systems to update functionality/
performance and to mitigate technology
obsolescence problems.
• Design refreshes must balance the cost avoidance
from mitigating obsolescence problems with
potentially large costs associated with system
redesign and re-qualification.
Mitigation of Obsolescence
Cost Analysis (MOCA) Tool
4
µ actual
µ forecast
98 9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9 10 11 12
3
•
•
•
•
•
-0.23
σ = 3.1M
2
40
50
60
70
DRAM Size (Mb)
1
0
0
2
4
6
8
10
DRAM Size (Mb)
LOT1
LOT2
12
14
16
18
2 boxes (4 levels of hierarchy)
20 boards (13 unique)
1310 parts (175 unique)
20 year sustainment life
Scheduled manufacturing
during first 12 years
• All obsolescence forecasts
from TACTrac data (circa
1998)
Actual Production Plan
LOT3
Actual budgeted refresh
LOT4
determined from the state of actual
LOT5
obsolescence events and production
LOT6
LOT7
LOT8
LOT9
LOT10
LOT11
LOT12
Predicted refreshes
from circa 1998
MOCA analysis
MOCA Forecasted Refreshes
98 9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9 10 11 12
Sponsors Include: U.S. Air Force, Lockheed Martin, Northrop Grumman, Naval Surface Warfare Center (Crane), PartMiner, CALCE Consortium