Why not! I’m
French, grew up by the sea in the Indian ocean and studied around to get my
French and a bit of American University education. I’m now approaching an age
of relative maturity (36) and have a 5-year-old son teasing me about my grey
hairs that start to appear... I’ve been working in the Real Time Simulator
Development domain for 8 years now. My very first task at the Experimental
Centre was to introduce the notion of distributed system design to the EEC’s
Real Time Simulator, which as I am sure we will address later, was a very
interesting challenge at that time.
I think
EUROCONTROL have always been champions of TLA’s (three letter acronyms) and ACE
is another fine example. ACE in fact stands for ‘AVENUE-compliant
ESCAPE’ and as you correctly calculated, is the 7th
generation of facilities at the EEC. This new generation of ATM real-time
simulator has been developed within the EATMP framework by the ERIS programme
from September 2001 to September 2002. As its acronym indicates, the ACE
platform is built upon a previously released and much abused version of the ESCAPE
platform, the main evolution being the adoption of the AVENUE architecture.
This standard architecture consists not only on specific logical module
decomposition but also imposes the use of dedicated API (Application Programme
Interface expressed in an OMG-based standard language called CORBA IDL).
ERIS – for me, without any doubt, you’re facing here the trickiest acronym at the EEC!!... Just remember ERIS as the host programme for the production of the EEC ATC simulation facility.
ESCAPE – EUROCONTROL
Simulation Capability And Platform for Experimentation,
the EEC’s first open system that served as a foundation to subsequent
activities.
AVENUE – An ATM
Validation Environment for Use towards EATMS
OMG – Object
Management Group, the standardisation body for open systems in a multitude of
domains, including transport, formed by a very large number of major companies.
To
summarize, AVENUE is a standardised architecture for ATM validation platforms.
‘Standard’ in this context means that it is a commonly agreed view among
several key actors (ATM Supply Industry, ATSP and R&D establishments)
within the ATM domain.
The best
analogy would be that of a box of Lego bricks…. They are of different colours,
forms and sizes, but due to the common interface, can be assembled into many
different toys…. AVENUE, simplistically viewed, provides a number of standard
bricks (logical modules) and the interfaces that allow them to be assembled
together to provide different levels of ATM functionality….. approaching
“plug-and-play”, a term that many fear!
To give a
brief history, the AVENUE project was launched in 1997 by the European
Commission (EC) in the context of the 5th Framework Programme,
following work performed in the 4th Framework under the name of
PATIO. A consortia comprising key European partners drawn from Research, Air
Traffic Service Provision, the ATM supply industries and the EEC worked
together to achieve the identified goals until April 2001 when the project was
completed. The main objective of the AVENUE project was to deliver a logical
architecture (supported by a set of defined APIs) for a platform capable of
validating new concepts and functionalities for ATM in both simulation and real-world
demonstration. In addition to the architecture definition, the EEC and the ATM
supply industries supplied one or several components that were successfully
assembled into a demonstration platform, thereby building confidence that the
resulting architecture was not only paper exercise, but also could actually
support ATM simulations and evaluations.
First, I
would say that I’ve always seen AVENUE as the next natural step in the
simulator evolution. When we look back at the more recent simulator development
history, we moved from a monolithic, big-blob SIM5+ platform into an open
architectured, CORBA-based ESCAPE platform in 1997 to combat the limitations
inherent to architectures of that nature.
Although an
excellent tool for the HMI and behavioural flexibility necessary in supporting
the multitude of different requirements from the customers of organisational
simulations, we had little to no functional flexibility within SIM5+, into
which the introduction of new functionality required major work, disturbing
large proportions of the global platform software, proving costly, particularly
time consuming and inefficient. Its modularity was quite poor but equally, we
relied on proprietary communication protocols, parts that were almost always
affected by the changes.
At that time
there were visionaries at the EEC who recognised the necessity to evolve the
system to facilitate the notional introduction of “plug-and-play” functions
into a common architectural framework. This coincided the EATCHIP III period,
where Operational Requirements Documents were being developed for a number of
“advanced” functions. To ensure that these new concepts be developed and
integrated into the centre’s simulation facilities, it was essential to create
an adequately open and distributed system favouring the notion of
“plug-and-play”.
Hence the
birth of ESCAPE. We were finally able to enrich the platform more efficiently
than before, with less disturbance to the overall system, and through the
introduction of the CORBA standard communication protocol, were able to focus
software development on the ATM domain, simplifying both system and integration
aspects. This was leading edge at the time, especially in ATM and although we
had come up with a nicely distributed system with modular ATM functionality,
our ATM data model and general platform architecture (service interfaces) were
still largely based upon the legacy system. CORBA had obviously brought
standardisation to our architecture but certainly not to the ATM domain side.
ESCAPE was still a proprietary, CORBAised version of SIM5+….
Absolutely.
Although the AVENUE data model and platform architecture is built upon CORBA,
it differs from ESCAPE in its recognition (as commonly agreed pre-standard)
from a large number of key players in the European ATM community. We can now
claim that our platform relies on standards in both domains, technically (with
a CORBA middleware) and functionally (with the AVENUE data model and service
interfaces). In fact, it is now sited as a reference or foundation for formalised
standardisation with the European Standardisation bodies….
I would split the benefits into 2 categories. Firstly, by providing and maintaining this commonly agreed architectural foundation, it strengthens EUROCONTROL’s supporting role to the European Commission’s objectives in the areas of interoperability and standardisation. This standardisation will allow the controlled evolution of the foundation in close partnership with the ATM supply industry, providing a common view of the ATM system architecture, which will ease development and integration of future ATM products and components. The adoption by the supply industry of this foundation will greatly accelerate the transfer of products of research into operations and contribute largely to a more flexible and progressive evolution of the ATM system.
Secondly,
the benefits for the ERIS programme itself. This commonly agreed framework
favours industrial partnerships, allowing ERIS to move away from pure software
development and to re-focus its activities upon the ATM research domain, whilst
providing the necessary support to international standardisation.
At first
sight, you may think that architecture standardisation does not bring any
advantage to the platform users… Effectively this should be transparent to the
end user.
For many
years, the EEC suffered from an image of a “dream factory” by the supply
industry. This had both positive and negative connotations attached… On the
negative side, many users found very promising ideas at the EEC and assumed
that these were readily available, off-the-shelf from within the supply
industry. As these were prototypes built upon in-house solutions they were
often frustrated.
This
situation will now change. With greater industrial participation, the users
will have better access to industry-based ATM products which provide
experimenters with both a credible and operationally proven foundation upon
which they can integrate new prototype functions in support to the concept
evolutions, something that was hard to envisage at the time of SIM5+….. Equally
beneficial is that the supply industry will have access to the new
functionality and therefore be in a position to prepare for their product
evolutions accordingly.
To bring one
example to the forefront, ACE2004A (the version of ACE to be delivered at the
end of 2003) will include an advanced trajectory predictor (TP) from INDRA that
will enable, for example, advanced AMAN experimentations with controller time
input orders.
Compared to
current ESCAPE platform, we have developed ACE to maximise the re-use of
existing hardware when migrating from ESCAPE. Indeed, all pilot (PWP) and
controller (CWP) working positions, that is to say, a large proportion of the
hardware composing an ACE installation, will be kept as ACE will run these on
either HP-UX 11, SUN-Solaris 8 or PC-Linux. However, using ACE will have an
impact on the GROUND system hardware since its applications (FDPS, RDPS, ATC
tools, datalink and Supervision) can be run exclusively under SUN-Solaris 8.
Vacation
maybe, sometimes in 2005!....
For more
information, click on this link to see a brochure
about ACE (warning 4 MB).
Or contact Anthony Inard.