The Future of Embedded Systems: Are you ready to transform your embedded development. May 1, 2014 webinar slide deck.

1. The Future of Embedded Systems
Are you ready to transform your embedded development?
Thursday, May 1st

2. 2
Moderator and Panelists
Moderator:
Jim McLeod-Warrick,
Managing Partner, Beacon Technology Partners LLC
Jay Andersen,
VP of Professional Services, Wind River
Rommel,
VP, VDC Research
Ken Caviasca,
GM, Intel IoT Solutions Group,
Platform Enabling and Development (PED)

3. 3
Methodology
 Web survey
• 674 usable responses (representing a confidence interval of +/- 3.7%)
• North America = 392 completed responses
• EMEA = 135 completed responses
• Japan = 30 completed responses
• Rest of Asia = 117 completed responses
• Fieldwork dates: October 10th
through November 27th
, 2013
• Sample provided by: UBM Tech
• Respondents screened for:
– Involvement in developing embedded applications and primary focused on software/firmware
• Research team: Beacon Technology Partners (www.beacontech.com)
– Methodology, fielding, data reporting

4. 4
Respondent Characteristics
Job Function %
Staff 75
Software design engineer 42
Systems engineer / Systems architect 16
Hardware design engineer 8
Hardware / Software engineer 3
Test, quality, manufacturing or operations
engineering
3
Other 3
Management 25
13
Engineering manager/director 11
1
Region %
58
20
17
5
Embedded Applications Focus %
59
41
Years of Experience
14.9
Size of Team
31
Sales Volume
$714.80Average Sales Volume (in millions)
Average Years of Experience
Balance between hardware and software
Japan
North America
Europe
Rest of Asia
Primarily software-oriented
Project leader
Corporate management
Average Number of Engineers

5. 5
Embedded design teams are more apt to be working on “new to the
world” projects as opposed to incremental improvements to existing
designs
1. What proportion of the projects you personally work on are:
Base = Total
New design = 58%
Existing design = 42%
New design = 57%
Existing design = 43%
New design = 51%
Existing design = 49%
New design = 57%
Existing design = 43%
Worldwide
New design = 57%
Existing design = 43%

6. 6
2. Approximately how many weeks did your [new design/upgrade of an existing design] project require, from initial documentation and exploration until the
project was deployed and maintained for end users?
3a. What proportion of time was required for each of the following stages in your [new design/upgrade of an existing design] project:
3b. Approximately how many software or firmware engineers were directly involved at this stage of the design of your [new design/upgrade of an existing
design] project?
The “develop” stage – whether for new or existing design projects
– usurps the largest proportion of a project’s timeline
Average Weeks
Base = TotalStage of Design
Indicates significant difference
at 95% confidence interval
between regions
Total Weeks
New design = 51.7 weeks
Existing design = 29.6 weeks

7. 7
Question
Typical embedded software design requires
nearly a year to complete (for new projects) and
30 weeks for revisions (to existing designs).
How does that compare to the past?

8. 8
Most embedded projects are delivered 3 to 5 months late
3c. Was your [new design/upgrade of an existing design] project delivered…?
Project Timeline
Base = Total
Indicates significant difference
at 95% confidence interval
between groups
New
design
Existing
design
% %
less than 2 months 26 26
2 to less than 6 months 15 12
6 to less than 12 months 8 3
12 to less than 18 months 3 1
18 months of more 2 1
Cancelled 1 2
Avg months late 5.1 3.5
Late by…
Amplitude of Delay
North American engineers
more apt to be “on schedule”

9. 9
Question
Roughly half of all embedded software projects
are delivered three to six months late. Why
don’t we do a better job anticipating these
delays?

10. 10
Internal challenges (e.g. murky requirements and overly tight timelines) and
external forces (e.g. hardware availability, system interconnections) combine to
hinder time-to-completion goals
4. We’d like to know more about the various factors you confronted during your [new design/upgrade of an existing design] project.
Mean share of project problems
Base = Total
Mean Scores
Share of problems arising
from:
Internal issues = 63.5
External Issues = 36.5
Asian engineers under more pressure
than other embedded developers
to reduce costs

11. 11
Question
The chief reasons for delay, according to our survey, are “internal” –
primarily due to changing or murky specs/requirements or unrealistic
schedules.
1.What internal or cultural changes might be implemented that can
shorten subsequent development time and reduce the possibilities of
project delay?
2.How can we do a better job of anticipating the unknown or “external”
hurdles such as technology changes, emerging industry standards or
new governmental regulations?
Q4: Assigned to Jay, Chris
Q6: Assigned to Jay

12. 12
Most agree that further investments in tools, training and support
would improve the ability to hasten time-to-completion
9b. Please tell us whether you agree or disagree with each of the following statement about your embedded software or firmware design activities.
“I feel that with right combination of tools, operating systems and consulting,
training and support, we could improve our development efficiency significantly”
78% Agree
Base = Total

13. 13
Although proof of ROI is still necessary, most are open to increased
spending to achieve a better embedded development environment
11. Please tell us whether you agree or disagree with each of the following statements about your embedded software or firmware design activities.
Agree Neutral Disagree
Management more
likely to feel ROI is
necessary before
investing in tools –
and are also more
likely to consider
using 3rd
parties to
help improve
development time-to-
market

14. 14
Base = Total
%
Right combination of tools 76
Willing to invest in 3rd party 49
Demonstrate clear ROI 60
Asian respondents are particularly open to this level of
investment in tools, training and support
% Agree Summary
Indicates significant difference at 95% confidence interval between regions
%
Right combination of tools 80
Willing to invest in 3rd party 53
Demonstrate clear ROI 51
%
Right combination of tools 89
Willing to invest in 3rd party 63
Demonstrate clear ROI 60
%
Right combination of tools 56
Willing to invest in 3rd party 48
Demonstrate clear ROI 67
Right combination of tools: “I feel that with right combination of tools, operating systems and
consulting, training and support, we could improve our development efficiency significantly”
Willing to invest in 3rd
party: “Our software group would be willing to invest if a 3rd
party could prove
how their solutions, tools, and services could significantly increase developer efficiency and improve time-
to-market our software development”
Demonstrate clear ROI: “Our company has to demonstrate clear ROI in order to invest in better
development solution, tools and services”

15. 15
Questions
Economic realities makes cycle time reductions difficult.
1.What resources do embedded project leaders have at their
disposal to surmount the challenge of fewer engineers available to
work on more projects?
2.What ROI metrics are used – or should be used – to help
development teams make better business cases for investing in
improved embedded development tools and processes?

16. 16
There is a fairly even split in preferred OS platform between
commercial and open source/“roll your own” solutions
5. Which of the following best describes your company’s embedded OS platform or scheduler use for most of its embedded applications?
Preferred embedded
OS platform or scheduler
Base = Total
Using Commercial
OS
NA = 44%
EMEA = 30%
Japan = 63%
ROA = 35%
Indicates significant difference
at 95% confidence interval
between regions

17. 17
Question
What accounts for the persistence of “roll your
own” or “do it yourself” open source OS
platforms – and do you see this changing in the
near future?

18. 18
Pros and cons of using commercial vs. open source
OS platforms
6a. Which of the following best describes why your company does not generally use a commercial embedded OS for your applications?
6b. Which of the following best describes why your company generally uses a commercially supported embedded OS for your applications?
Reasons for using commercial OS
Other = 8%
Base = Those who use commercial
embedded OS platform
Reasons for NOT using commercial OS
Other = 8%
Base = Those who use in-house
embedded OS platform

19. 19
Support and long-term relationships undergird commercial OS platform
preference, while short-term cost considerations drive open source OSs
11. Please tell us whether you agree or disagree with each of the following statements about your embedded software or firmware design activities.
% Agree
Indicates significant difference
at 95% confidence interval
between groups

20. 20
Question
The survey shows us that those who shy away from using commercially available 3rd
party OS platforms do so due to concerns about complexity and upfront cost.
1.What are some of the implicit or hidden costs – in both time and money - that
embedded developers should also be aware of when deciding whether to rely on open
source or internally developed OS platforms?
2.What might be the impact of the coming “Internet of Things” and the billions of
interconnected devices IoT entails on the use of various OS platforms?

21. 21
“Iterative” or “agile” development methods are used with
greater frequency than are traditional “waterfall” approaches
7. What development process methodology do you or does your software development team generally follow for most of its embedded development?
Development process methodology
Base = Total
Adopted Agile
NA = 33%
EMEA = 39%
Japan = 23%
ROA = 32%
Indicates significant difference
at 95% confidence interval
between regions

22. 22
Question
A good deal of common wisdom suggests that
“agile” or “iterative” development is displacing
“waterfall” software development methods. Do
you agree – and if so, why this change in
development methods?

23. 23
Demand for training and consulting services will help spur
the “shift left” movement in development cycles
11. Please tell us whether you agree or disagree with each of the following statements about your embedded software or firmware design activities.
Agree Neutral Disagree

24. 24
Question
What specific training and educational
investments can you suggest to software
development teams to improve their design
cycle times?

25. 25
Average Utility
*Attribute utilities scores are relative rather than
absolute; even the lowest-rated attribute has a
positive impact (albeit less important) on overall
interest in embedded development solutions
Represents
over 60% of
overall utility
Most persuasive solutions to reduce embedded development time-to-completion

26. 26
Question
Simulation has been suggested in the past as
enabling work-arounds to hardware-related
issues, yet early attempts at simulation proved
unfruitful. What recent advances can you cite in
system simulation that would give engineering
leaders more confidence?

27. 27
Substantial opportunities for future adoption of embedded development solutions,
particularly design services, system simulations and pre-integrated software
Just
starting
Along
the path Mature
8/8a. How interested are you in using each of these programming tools or practices as part of your embedded software development?
9/9a. How interested are you in each of the following embedded development tools or technologies?
Base = Total
Ranked by
total utility

28. 28
How to Read this Chart: The chart indexes the average reduction in development time due to adoption of various solutions.
Average
Ranked by potential impact
New design
Existing design
10a. Consider the following embedded development tools or technologies you previously evaluated. Which ONE do you feel would have the greatest
impact on your software development cycle time?
10b. Approximately what proportion of the total man-hours from during your [new to the world design/upgrade of an existing design] project might be
reduced if you or your development team were to adopt [INSERT ITEM FROM PREVIOUS QUESTION]?
Investment in these solutions (particularly for pre-integrated software solutions and high-level design services) are expected to
contribute to a 20-25% reduction in development cycles
Average potential reduction
in development time
New design = 11.9 weeks
Existing design = 7.2 weeks

29. 29
Question
While all of the listed solutions garner at least some interest,
several appear to be particularly suited to shortening development
cycles. In what specific and tangible ways might each of these
(1)shorten development cycles and
(2)meet and handle the external and internal issues which cause
schedule delays?

30. 30
Observations
 Although embedded development teams deploy considerable resources toward their designs, roughly half of
all projects are still delivered three to five months late
 Development teams devote 51.7 weeks to “new to the world” projects and 29.6 weeks for incremental improvements to existing designs
 North Americans devote more time to their projects than do other development teams elsewhere
 Chief challenges which hinder faster time-to-completion – changing or murky requirements; overly stringent
timelines; hardware unavailability; interconnectivity; and deluges of design defects
 Widespread interest in and demand for proven solutions to improve design cycle times, such as
 Utilization of fully supported commercial OS platforms
 Adoption of “agile” or “iterative” development methods, along with proactive debugging/testing
 Improved development tools and more fully functional IDEs
 Full system simulation and runtime solution stacks
 Better training
 Adoption of many of these solutions (particularly high-level design services, or pre-integrated software
specific to the application) promises to reduce overall embedded development time by 20 to 25% on average

31. 31
Audience
Q & A

32. 32
For more information:
Download Whitepaper:
http://www.windriver.com/whitepapers/transformation-time-for-e
Sponsored By:
Visit
www.windriver.com/embedded_transformation/