CE Linux Forum Activities

1
CE Linux ForumActivities

• Satoru Ueda
• CE Linux Forum
• Marketing Group Chair

2
CE Linux Forum start working well
Especially in Far East Asian region
CE Linux Forum
3
How about my idea? Even your small idea will be
welcomed. Lets release it to the forum, just as
you to release baby salmons to the upper stream
of river!
4
There may be your partners waiting to collaborate
with you.
5
You may be surprised some body will examine your
idea from the other aspects which you have never
been imagined. Some one will be happy to archive
them.
6
And lets use those solutions in the actual
engineering scene.
7
Some one may find the bugs
8
Some one may be inspired the other fantastic
technical idea!
9
Then, again, release them to the upper stream of
the river, Open Source.
10
All are empowered by the Open Source!
11
The Worst ScenarioWe used to be concerned
12
Just waiting something to be happened
13
Even you developed good technology but
not releasing to the community. May be you will
say I am too busy.
14
Building the fence to the world of Open
Source and prevent your engineers to
communicate with the outsiders.
15
The river will be dried up, fertile land will
become desert, and the Ocean of Open Source will
die.
16
Before CE Linux Forum

• December 2002, Panasonic and Sony unveiled the
  collaboration on Linux development for consumer
  electronics products

17
Establishment of CELF

• June 23, 2003
• Founder Panasonic, Sony
• About 50 member companies, now

18
First and the biggest mistake

• Released Standardized Specification
• April 2004 (Draft version)
• Just a SPEC without implementations
• Bashing from community
• Who to create the codes?
• We have learned that it was not ideal way to
  develop together with the community
• Implementation First!

19
Technical Conference in JapanWhat happened?

• Many non English speakers exists in the embedded
  Linux field
• No one dare to speak
• Need to break the Sound of silence
• Started Japan Technical Jamboree

20
Japan technical jamboree

• Not a technical conference but a jamboree
• Imagine boys and girls around the campfire
• Talking freely about their future, dream and past
  story
• In Japanese (a compromise)
• Speaking Japanese is far better than speaking
  nothing
• Started October 24, 2004 in Tokyo

21
Japan Technical Jamboree

• Getting to be exited from Fall 2005
• Decided to have bimonthly
• Increased the contributors

.Finally, I am a little embarrassed that my
code runs only on Kernel 2.4
May I give you a comment? Because your technology
seem so interesting, I have ported on latest
2.6.xx Kernel. May I demonstrate it right now?
22
Japan Technical Jamboree

• The venue for the next jamboree should be located
  near from Tokyo Tower!

???
Go beyond Linux Kernel
23
Japan Technical Jamboree
24
Japan Jam.Now Counted 19 times

• English translation volunteers
• To put the materials in Japanese into English
• Started the campaign
• Lets join the community!
• Encouraging Japanese developers to join CE Linux
  Forum and Linux Mainline international activities

25
CE Linux ForumInternational

• Worldwide Embedded Linux Conference
• Annual event from 2005 in Silicon Valley
• ELC 2008
• April 15th to 17th
• At The Computer History Museum
• Silicon Valley area, California
• Soliciting technical demonstrations

26
ELC 2008
Lets get together and talk technology!

• Not only embedded system developers
• Together with mainline community people
• Jonathan Corbet (LWN.net) admires
• ELC 2008
• From April 15th to 17th
• At the Computer History Museum

27
What is the True Value to Use Linux for Embedded
Systems?

• November 2007
• (Presentation at ET 2007)
• Jyunji Kondo
• Fujitsu Software Technologies Limited

This material is prepared by collaboration of
about 10 CELF enthusiasts.
translated by ikoma
28
Here is a Fact
Has it ever happend that programs running in home
appliances are published?
They have one thing in common - they all use Linux
?
Why do they release their source code?
29
Here is Another Fact
Memory Use Reduction when Using Dynamic Libraries
on Embedded Systems (by Yamamoto, Panasonic
Mizuyama, Panasonic Mobile)
Hints on Faster Booting Up (by Fukumoto, Sharp)
Linux Keitai Challenge to Better Performance (by
Fukunaga, NEC)
30
Here is Another Fact
Technologies presented here are what all software
(esp. embedded) developers want to know.
And they are all practical and have been proved
on real products
?
Why do they generously publish their own
achievements?
31
Here is Another Fact
It does not appear that GPL is the only reason
to force them to release their source code
The answer is
There is a True Value to use Linux
They are already aware of the True Value and
enjoying it
32
Then What is the True Valueto Use Linux?

• Key phrases are
• No need to build all yourself
• Ease of development
• More attractive products
• Maintainability

33
Agenda of This Presentation

1. The True Value explained using the key phrases
2. On so-called shortcomings of Linux
3. To make full use of the True Value of Linux

34
It is Linux, but ...
No need to build all yourself
Software
Applications
Middleware
Linux Kernel
Hardware
35
It is Linux, but ...
No need to build all yourself
Software
Apps
C Apps
Java Runtime Env
Misc Runtime Env(Python, Perl, Ruby, )
Java Apps
Misc Script Apps
Middle ware
Linbraries

• GNU Software

• Other FLOSS

Non FLOSS
Linux Kernel
Hardware
FLOSS Free/Libre Open Source Software
36
No Need to Build All Yourself
No need to build all yourself

• Background
• Embedded system software gets more complicated
  and larger
• Benefits
• Software (OS/Middleware) can be reused in Linux
• Assured quality
• Lower cost
• Shorter time to market

37
When Software Gets Larger ...
Ease of development

• It becomes a so complicated system that
  conventional RTOS can not support any more
• Can Linux features make it easier to devlop?
• Here are voices of converts from RTOS to Linux
• Wanted to protect memory with virtual memory
• Wanted adaptive scheduler

38
Wanted to Protect Memorywith Virtual Memory
Ease of development

• Large size has made bugs unavoidable
• If bugs remain in applications...
• Memory protection would prevent confusion as
  abnormal termination of the application
• Restarting the application would keep the service
  going on

39
Wanted Adaptive Scheduler
Ease of development

• Applications have been changed as they get larger
• Rich and CPU-hogging applications
• Applications supplied by third parties
• It is hard to determine scheduling order of all
  tasks at designing phase
• Want to delegate to adaptive control at runtime
• Loose coupling of components in the system

40
Making Software More Mature
Ease of development

• Another voice from develpers
• The more complicated a system is, the more it
  should be developed in open source way
• Post source code to the community as open source
  when to develop complicated software
• You will receive comments and/or fix proposals
  from experts in the open source community
• Collaboration with the world class
  super-developers
• Your published feature may return as much better
  one

41
Case Study at CELF (1)
Ease of development

• prelink
• Technology for faster start up of application was
  presented
• It partially depended on a CPU architecture
• A developer called for information on unsupported
  CPU
• Developers including prelink experts, CPU
  venders, started to collaborate
• Expected feature was realized in a short term
  with the help of collaborators

42
Case Study at CELF (2)
Ease of development

• Toolchain
• A developer found a bug in a toolchain
• He fixed and released at CELF, but it was for the
  older version of Linux
• When he was to support the latest version...
• Another developer has already ported it
• One fix produced two fixed versions
• When sharable, collaboration is productive

43
Higher Productivity
Ease of development

• Some developers regard it as merit of Linux that
  high level trouble shooting and tracing
  functions are incorporated
• Originated from mission critical field
• Not available on ordinary RTOS for embedded
  systems
• Helpful when verifying large scale software

44
Hiring/Educating Engineers
Ease of development

• In proportion to the size of system, the number
  of necessary developers also grows
• How to get excellent developers involved is
  one of keys for successful project
• Merit of Linux
• Low barrier from outside of embedded world
• Ease of finding development partners
• No need to relearn development methods

45
Why Software has Become Larger?
More attractive products

• To satisfy users needs with more attractive
  products
• Boosted with dramatic progress of semiconducter
  technologies

46
More Attractive Products
More attractive products

• Linux is one of the operating systems where the
  latest technologies are first supported
• Support of new services and devices
• Kernel which keep evolving
• Linux is evolution, not intelligent design
• Device vendors supply device drivers
• Also linuxdriverproject.org launched
• Developer communities implement standards
• IPv6, DLNA, SIP, SMB, Bluetooth, OBEX, SNMP, ...
  

47
Maintainable
Maintainability

• Characteristics described above also improve
  maintainability
• Memory protection with virtual memory
• High level trouble shooting and tracing functions
  incorporated
• Open source makes it possible to keep maintained
  continuously

48
Intermediate Summary
Maintainability
Product Life Cycle Product Life Cycle Product Life Cycle Product Life Cycle QCD QCD QCD
Planning Design/Manufacturing Sales Maintenance Quality Cost Time to Market
No need to build all yourself
Ease of development
More attractive products
Maintainability
49
Often Heard Rumors on Linux
Shortcomings of Linux?

• Linux cant do realtime processing
• Linux is a memory hog
• Linux requires high spec hardware
• Reality check!

?
50
Linux Cant Do Realtime?
Shortcomings of Linux?

• Conventional RTOS achieves realtime performance
  of less than several microseconds
• Right, Linux can not achieve this level yet

• But recent Linux is
• getting much better
• Improved up to dozens
• of microsecond level

51
Linux is a Memory Hog?
Shortcomings of Linux?

• Linux as OS itself spends more memory than RTOS
• But it is application that spends most of memory
• in a large system
• Of course, Linux is
• further evolving
• For products with
• hard disk drive,
• memory restriction
• is not that severe

52
Linux Requires High Spec Hardware?
Shortcomings of Linux?

• It is true that Linux does not run on 8bit or
  16bit microcontrollers
• But to do a complicated processing, hardware as
  such is required
• Low spec CPUs
• without MMU are
• also supported

53
To Make Full Use of True Value of Linux
Software
Software
Apps
Java Runtime Env
Misc Runtime Env(Python, Perl, Ruby, )
C Apps
Java Apps
Misc Script Apps
Middle ware
Libraries

• GNU Software
• Other FLOSS

NonFLOSS
Linux Kernel
Hardware
54
To Make Full Use of True Value of Linux
Of course you can use as is ...
Java Apps
Misc Script Apps
But if you work with experts...
Software
Apps
C Apps
Java Apps
Misc Script Apps
Middle ware
Linux Kernel
Hardware
55
To Make Full Use of True Value of Linux

• CELF is a good starting point
• Variety of specialists are participating
• Full of communication tools available
• Wiki
• Mailing lists
• Jamborees,
• Embedded Linux Conf.

Low
Communication density
High
56
CELF Wiki, a Treasure Trove
57

• Linux accelerates to build your embedded systems
• And it is ready

58
Thanks!