9:00 – 9:30: Introductions & Overview of the RT&ES
Forum
(Joe Bergmann, RT&ES Forum Director)
Joe started
off the Tuesday meetings with an introduction to the RT&ES Forum and
their focus on “Dependability through Assuredness™”. He
highlighted the Working Groups: Open Architecture for RT&ES, IEEE
POSIX Standards, Profiles & Certification, Security & Safety-Critical,
Safety-Critical RT Java/JSR 302, and Secure Mobile Architecture.
Joe went over the Agenda and laid out expectations for the week.
The
presentation for this session can be found here.
9:30 – 10:00: Dependability through Assuredness & AADL – A
Perspective from Japan
(Jack Fujieda, CEO & President, REGIS, Inc.,
Representative and Chairman of The Open Group, Japan)
Jack provided great
insight into the mindset of the Japanese business community,
highlighting the negative effect of a high “change-over” rate for
top executives and business strategists, which reportedly happens about
every three years, so the business leadership is always changing very
fast, before they can carry out the objectives of the previous business
team. That is why new
venture, or forward-looking companies, who are focused on identifying
major problems and improving the company, can have a very positive
effect. But Jack emphasized that it is important to have low-hanging
fruit to show productivity. With the quick change-over, short-term goals
and quick wins are more important than long-term strategy and promises.
Forward-looking standards organizations or venture companies must
convince Line Executives that they can provide quick wins. Line
Executives are not quick to change – they are very proud of what they
have done. And to them the bottom line is the primary criteria – so
reducing costs is the most important thing - rather then investment in
the future. Jack Fujieda is on the
Dependability Task Force started under
the Ministry of Economy, Trade and
Industry (METI) in Japan and they
are working on solutions for Dependability through Assurance.
He emphasized the importance of providing embedded systems that
include the necessary security and reliability controls,
which may lead to larger multi-processor systems.
Jack provided a great presentation on the statistical analysis of
cost factors in implementing Discovery and Fault systems, and the effect
of implementing fault tolerance and discovery during the requirements
and design phases of a project, where architecture is based on a
Dependability Meta Model.
Fujieda-san
briefly covered AADL, and emphasized that there was a component that
focused on human policies for different situations where humans are
responsible for faults, and that AADL provides a method for introducing
policies to mitigate the human risks. The Dependability Task Force
proposed the AADL Error Model as a way to formalize operations as a
recommendation to the government. But there are more studies necessary
to validate the recommendation, which is dependent on more money from
the Japanese government to study AADL.
The
presentation for this session can be found here.
10:00 –
10:30: AADL Overview
(Ed Roberts, Elparazim)
Ed Roberts presented
an overview of AADL as a prelude to
the following session where he offered a more in-depth tutorial.
AADL comes out of the Society of Automated Engineers AS
5506A SAE International Standard document AS 5506A. AADL is a formal
modeling language for describing software and hardware architecture. He
stressed that AADL requires starting at the design phase – and
reiterated what Jack Fujeida had presented in his statistical analysis,
that if you can anticipate the faults and risks at design time you can
eliminate about 70% of the problems.
The focus of AADL is on physical systems: state machines,
busses, processes, and is targeted toward embedded real-time systems. Ed
emphasized that there is work being done to “merge” SYSML and AADL,
which involves the profile MARTE as a basis of conversion. MARTE is a
profile that deals with non-functional elements (e.g., size of a
footprint).
The presentation for this session can be found
here.
11:00 –
12:30: AADL Tutorial and Discussion
(Ed Roberts, Elparazim)
Ed Roberts provided a great AADL tutorial, going into
detail on how to use AADL for the following component categories:
Software, Execution Platform, and System Components, providing a detailed
view of the language.
The presentation for this session can be found
here.
14:00 – 15:30: TOGAF for Newbies – Tutorial
(Judith Jones,
Architecting-the-Enterprise)
Judith Jones provided a fabulous overview of the TOGAF
approach and the ADM, continually bringing it back to the Real-Time and
Embedded Systems environment and where in the ADM high-assurance
architects might need to consider what roles, inputs, outputs,
artifacts, and next steps would typically be specific to the RT&ES
“world”. Judith
suggested that when the RT&ES Forum looked in more detail at these areas
and worked through the real-time-specific aspects, that we should
consider as a next step coming up with a Reference Model that would be
useful for all high-assurance architected solutions.
The presentation for this session can be found
here.
16:00 – 17:30: Secure Mobile Architecture
(Joint with Security Forum)
Refer
to the Security Forum report on this session.
TOGAG-Real-Time
Extensions: Ed Roberts and Edwin
Lee with input from Judith Jones will be driving this work with the RT&ES
Forum, advancing the TOGAF-RT&ES Reference Model work that was begun
during the last conference and mapping the RT&ES-specific
requirements to the TOGAF ADM. See the report on the
Thursday session for additional actions related to this area.
Modeling
Language Tutorials: The feedback
on the AADL tutorial was very positive, and it was suggested that we
allow for an additional tutorial from Ed Roberts on SYSML either via
WebEx before the Boston Conference or during The Open Group Boston
Conference (July 19-23, 2010).
Secure
Mobile Architecture: Follow-up
actions are for the Forum Directors to encourage return of completed
use-cases, so members can review and derive common requirements across
these use-cases and from that work out how to organize our architectural
objectives. We will also check on options to add further relevant
use-cases to this study. We will plan to run a project conference
call every two weeks (Steve will advise the best day/time for these
calls) to progress return and review of use-cases and assure shared
understandings on deriving common requirements, and plan to run a
half-day SMA project workshop in our next Conference (in Boston.)
