Research
Without realizing it, many of the technologies that we use on a daily basis are in fact based on standards that are subject to a broad consensus generally on a global level : we cannot imagine exchanging documents via the Internet without having recourse to common conventions, such as Internet (IP) routing protocols or web presentation formats (HTML or XML). How can we share these common resources via telecommunications networks or IT grids ? This special feature underline INRIA research teams who have contributed to these standards with several industrial and academic partners within international organization for standardization and normalization.
We cannot imagine exchanging documents via the Internet without having recourse
to common conventions, such as Internet (IP) routing protocols or Web
presentation formats (HTML or XML). How can we share these common resources via
telecommunications networks or IT grids?
In the IT world, rapid changes in technologies and the desire on the part of companies
to provide products integrating new solutions in a timely fashion have led, over
the past two decades, to the creation of specialised consortia, often unincorporated,
uniting players from public and private research. These committees issue
recommendations, that is, a set of specifications describing the implementation
principle for the technique with a view to interoperability between the products
manufactured by the players. Each company is free to follow these recommendations
or not, but generally they are adopted and thus become a standard. The term standardization
therefore implies the adoption, by most of the participants, of specifications
drawn up within a collective working framework. All the same time, the players
are aware of the advantage of being involved in the activities of standardization
organisations (ISO, ETSI, etc.) that issue standards that are more regulatory
in nature.
At INRIA (National Institute for Research in Computer Science and Automatic Control),
for a number of reasons, this standardization culture has quickly entered into
the everyday practices of a number of research teams. For these researchers, it
is sometimes the best way to distribute and sustain their solutions. For some
time now, the Institute itself has travelled this path by hosting the W3C (World
wide web consortium), the consortium responsible for the development and promotion
of Web standards, from as early as its creation in 1994 up to 2002. In this way,
some 20 INRIA members participate in the W3C technical team. Moreover, an INRIA
researcher, Jean-François Abramatic, presided over the W3C for 4 years,
up to 2001.
An approach unequivocally supported by the Institute
INRIA solidly supports the standardization approach, says Gérard Giraudon,
Director of Development and industrial relations: « In certain sectors,
standardization is the only way to widely disseminate the work of researchers.
For example, in the world of networks, irrespective of the merits of a technical
solution, a solution will never be adopted by economic players if it is not recognised
as a standard. Standardization is therefore indispensable to ensure the future
existence of these technologies. It also offers worldwide scientific recognition
of the work of researchers and in doing so, raises the profile of the Institute. This
comes hand in hand with economic benefits: companies dedicated to the growth of
technological developments and reinforced cooperation with the industrial companies
concerned ».
Assisting researchers who wish to promote their solutions by providing financial
and human resources is in particular indispensable in the telecommunications and
middleware domains where the scientific, economic and social stakes are enormous.
A number of spin-off benefits
It is clear that the determination on the part of the institute and scientists
conducting INRIA research projects, manifested by their efforts on these standardization
committees, has contributed to international recognition of French research in
STIC. These organisations provide moreover both a vehicle where researchers can
showcase their solutions and also an area for testing, which often enables participants
to anticipate technical and scientific developments. For participants, in addition
to being an expert in a field, it also means having the opportunity to discover
new fields of research, which in turn provides a way of monitoring technological
developments.
At times, spin-off benefits can be as far-reaching as the creation of innovative
companies, such as UDcast, which was founded as a result of work conducted by
a research project to standardise the integration of satellite links in the Internet
infrastructure (UDLR protocol, RFC 3077 at IETF) and Luceor, which is based on
the OLSR protocol (RFC at IETF). In other cases, industrial partnerships and joint
projects are initiated or reinforced such as within the framework of the Carroll
research programme uniting INRIA, Thalès and the CEA in the software engineering
and middleware technological fields.
Afnor - Standarmedia: an inventory of the standardization committees in the TIC field: http://www.standarmedia.com/
The protocols which are the keystone of the Internet must be adapted to face
the explosive growth in numbers of connected machines. In addition, the components
that integrate these protocols must be reliable, capable of high performance
and compatible with one another.
We all know the success of the Internet that started at the beginning of the
1990s. The deployment of the Internet allowed access to numerous new online
services and applications. The price to pay for such enthusiasm was that the
limits of the worldwide network were soon brought to light. The Internet protocols,
a set of rules, conventions and mechanisms that make it possible for the network
to work correctly, are on the front line. These protocols require in particular
that each machine be allocated an address—its IP address—an indispensable
open sesame prior to all network communication. The IP address is a number of
fixed size, and the number of connected machines is now exceeding the maximum
allocation capacity (slightly more than 4 billion distinct addresses).
At the
present time, the network primarily uses the IPv4 protocol (Internet Protocol
version 4) that was standardized by the IETF (Internet Engineering Task Force),
the Internet protocol standardization body. However, just like phone numbers
had to go from 8 to 10 digits in order to increase the connection capabilities
of the French telephone network, a new addressing mechanism allowing for a larger
number of Internet addresses is currently being deployed. This mechanism called
IPv6 should soon become the standard used by the new Internet. The first IETF
specifications were published in December 1995.
Very early on, the major French players regrouped within the G6 association
to foster the development and deployment of the new version. The ARMOR team
(Rennes) was very active in it from the beginning. The team gathers together
researchers from a department of ENST Bretagne (National Superior School of
Telecommunications of Brittany) specializing in network protocols and researchers
from INRIA Rennes well-known for their competence in network modeling, test
and evaluation.
Preparing the Internet of the future
The team proposed a solution to the addressing problem at the IETF. The solution
involves a mechanism that can be used temporarily to solve the lack of IPv4
Internet addresses. Moreover, much of the team's work concerns the transition
from IPv4 to IPv6, via the temporary allocation of IPv4 addresses only when
an application requires them. This makes it possible to use the new version
of the protocol whenever possible and the “old” applications can
nonetheless benefit from the advantages of IPv6 (such as mobility, self-configuring).
ARMOR research scientists proposed solution for the protocol itself and its
impact on applications, such as the DSTM technology (Dual Stack Transition Mechanism)
which was submitted at the IETF in 1999 to let IPv4 and IPv6 coexist on the
same machine. DSTM is in competition with other solutions. Three implementations
have already been effected, at ENST Bretagne, by the ETRI (Electronics and Telecommunications
Research Institute) in Korea and by Hewlett Packard.
In addition, problems that were not addressed by IPv4, such as router configuration
(routers are the communication nodes of the network), are likely to take on
increasing importance with IPv6, due to the explosion in numbers of technologies
involved, machines connected to the network, and growing complexity thereof,
since the network topology changes with user mobility. Concretely, if it is
now possible to automatically configure machines, this is not the case for routers,
which require the intervention of a network administrator. The latter has to
configure the various links and routers of the network, and renew the configuration
every time the network topology changes. This problem occurs in particular on
the local scale, when a small company needs to deploy an internal network, or
in the home where various wireless technologies have the wind in their sails.
In 2002, researchers of projects ARMOR and ARES (Lyon) proposed a protocol?l called
NAP (No Administration Protocol) at the IETF, for the self-configuration of
IPv6 routers and network. It was the first proposal on the subject. It did not
however attract enough attention at that time to warrant the creation of a dedicated
work group. The researchers nonetheless continued their work in the framework
of research projects with France Telecom and Alcatel. There is little doubt
that the topic will soon be hot again at the IETF.
Testing the interoperability of new components
In parallel, in order to ensure a reliable deployment of the network, it is
crucial to check that the new IPv6 routers conform to the specifications already
defined by the IETF and are compatible with one another, even coming from different
manufacturers (Cisco, Hewlett Packard, Samsung,...). To achieve this goal, each
new product must be tested for compatibility with the specifications and interoperability.
This is one of the activities of ARMOR research scientists who have been developing
protocol testing methods for four years in collaboration with the ETSI (European
Telecommunications Standards Institute), the European standardization body in
matters of telecommunications. Comparable testing processes developed by American
and Japanese departments have been implemented in the United States and in Asia.
In order to ensure an international validation process, called the IPv6 ready logo program, the different players regrouped last year at the instigation of
the IPv6 Forum, the international organization promoting IPv6. The program is
based on three test sessions a year, successively in Japan, the United States
and Europe, during which all the companies involved in IPv6 deployment have
their components tested. ARMOR was instrumental in launching this worldwide
certification program and is its European representative. Eight researchers
and engineers are working on it today around César Viho. About a hundred
IPv6 components coming mostly from American (Cisco, Microsoft,...) and Asian
companies (Samsung...) have obtained this certification. In Europe, the first
certified companies are 6wind (FR) and Ericsson (SE).
Several INRIA teams are developing solutions adapted to new Internet uses while
ensuring exchange security.
The Internet has been constantly changing since the beginning of the 1990s and
its incredible worldwide development. It must continuously face new challenges
due to evolutions in the way it is implemented. One of the most worrying of
these challenges today is communication security: it relatively easy to spoof
the identity of a user and retrieve information in his or her name. Much research
is devoted to plugging this hole.
Another concern has assumed growing importance in recent years: the adaptation
of the network to ceaselessly growing user mobility. Like cell phones, computers
also are becoming more and more mobile. The challenge is to maintain their Internet
connection during any kind of journey, either inside the usual connecting network
for the device (also called its mother network) or over very large distances,
from one country to the next for example. With third generation cell phones,
billions of mobile devices will be connected to the network. Mobility also poses
specific security problems.
These different aspects are the object of standardization proposals at the IETF
(Internet Engineering Task Force), the Internet protocol standardization body,
especially in the context of the new IPv6 Internet protocol. Like its predecessor
IPv4, this protocol automatically allocates addresses to each machine (the IP
address is a long list of numbers), an indispensable open sesame prior to all
network communication. In order for users not to have to memorize the IP addresses
of the machines, a name is given to each of them. The correspondence between
machine addresses and names is stored in a large database called DNS (Domain
Name System).
Improving transmission security
However, access to this base is not secure, and for the time being, the identity
of each machine is not authenticated. An identity can thus easily be usurped
during a connection request and data can be hacked. Researchers from project
ARMOR in Rennes have been participating in an IETF work group called DNSext
on the topic since 2002. They are proposing several methods to make IP address
requests secure.
The IETF is also thinking about a scheme to make Internet communications secure.
Researchers from project PLANETE in Grenoble are defending a solution of cryptographic
IPv6 addresses called CGA (Cryptographically Generated Addresses), a secure
identifier, together with Sun Microsystems researchers. The solution makes it
possible for a machine to prove that it is using an address that was allocated
to it. There are many applications. The address spoofing problem can thus be
solved, as well as the IPv6 mobile connection highjacking problem. The protocol
used by the machine to configure its IPv6 address can also be made secure. Such
cryptographic addresses are in the process of being standardized by the IET.
The ensuing security mechanism is called HIP (Host Identity Protocol). It introduces
a new naming space to securely identify the extremities, i.e. the terminals,
in a communication. Concretely, each extremity or machine is allocated a secure
identifier obtained from its public key that will then be used by applications
in order to identify the extremities during a communication. In this way, such
upper layers become independent from the IP (v4 or v6) addresses, and thus from
the localization, and use secure identifiers. All this is made possible by the
HIP protocol that performs the conversion between identifiers and IP addresses.
Research scientists from project RESO (INRIA Rhône Alpes) are participating
in this work in the framework of the HIP work group, in collaboration with Sun
Microsystems. Project RESO is studying data transmission solutions adapted to
computing grids in which hundreds or thousands of computers are pooled together
over the network to supply large computing capacities. Security problems are
at the forefront of the team's concerns for obvious reasons of confidentiality
and protection of interconnected resources. Several of the team's proposals
have been accepted and are on the way to being standardized: one of them consists
in creating an extension of the DNS database for the HIP protocol, another one
in bypassing the DNS. Two implementations have already been completed, one at
INRIA and the other one at HIIT, a Finnish research department.
Taking mobility into account
Concerning mobility, the solution currently favored by the IETF is a protocol
called Mobile IP. With Mobile IP, mobile devices have a permanent IP address
known to all, and a temporary address in connection with its displacements.
All outside communications arrive at the permanent address and are then forwarded
to the temporary address. All these exchanges increase the risk of attack through
hacking of the signalization messages. Since the beginning of 2000, research
scientists of project ARMOR, with contributions from PLANETE, have been supporting
a solution to strengthen the security of communications between a mobile device
and its mother network at the IETF. This solution makes it possible for a device
using the Mobile IPv6 protocol to move without unveiling its permanent IPv6
address. The ARMOR proposition was accepted and standardized since June 2004
under RFC 3776 (Request For Comments). Research and implementation work are
continuing.
Another problem linked to Mobile IP protocols is that they process micro and
macro-mobility in the same way. A machine must communicate its new temporary
address every time it moves, irrespective of how far it moved, even though the
majority of displacements are local. Obviously, the resulting quantity of messages
generated is likely to crash the network. Researchers from project PLANETE are
proposing to adopt a hierarchical approach to the problem. The idea is to maintain
the principle of communication with the mother network and the Mobile IP protocol
for large displacements, but to manage local mobility without systematically
sending the information back to the mother network. An IETF work group was created
on the topic in 2000. Among the various solutions proposed, INRIA's was accepted
and is now defended by Ericsson. This solution is called HMIPv6 (Hierarchical
Mobile IPv6). It uses an internal protocol for local movements that hides them
from other users. In august 2005, the solution achieve the experimental RFC
status 4140. nother characteristic of the HMIPv6 protocol is that it makes it
possible to hide the geographical position of Internet mobile devices. As a
matter of fact, it only reveals a global address that supplies very little information
on the geographical location of the device. This is a sometimes useful feature.
This is the story of researchers who solved the problem of integrating satellites
into the Internet infrastructure, created a startup and are now developing new
applications of their standard.
In 1996, Eutelsat, a worldwide operator in satellite infrastructure, approached
INRIA in order to study the possibility of integrating satellites into the Internet
and thus transmit broadband data. Their idea was simple: geostationary satellites
put into orbit in view of the deployment of digital TV (using the DVB standard,
Digital Video Broadcast) have a huge unused transmission capacity. Why not try
and use them just like the fiber optics infrastructure to transmit broadband
Internet data (at that time the most powerful modems connected to the telephone
network were liberally transmitting 56 kilobits per second)? INRIA research
scientists rapidly identified the technological deadlocks—the routing
protocols used to transmit the information and data. These protocols work through
an exchange of data between routers—the communication nodes—that
let them know about the network topology and find transmission paths. This functioning
mode assumes a two-way communication: point B is assumed to be reachable by
router A if router A is receiving information from B.
Now in the case of a satellite, due to economic reasons, a home user may only
receive data: reception antennas are cheap as opposed to transmitter antennas
which are very costly. The idea was thus to use satellites to receive broadband
data and keep the telephone network and modems to send data, without modifying
the functioning of the network. In other words, the point was to integrate satellites
using one-way communications into a network, the routing protocols of which
assume two-way links.
A routing protocol dedicated to satellites
Eutelsat was convinced that satellites were an interesting solution for broadband
data transmission. The company thus funded a study carried out by the team of
project RODEO of Sophia Antipolis. A transmitter antenna was purchased and Eutelsat
set aside hours of satellite transmission for INRIA during a few years to perform
experiments. Within six months, the researchers had devised a first, theoretical
solution. The trick was to make believe that the satellite link was two-way.
In December 1996, INRIA created a dedicated work group at the IETF (Internet
Engineering Task Force), the Internet protocol standardization body. The group
was called UDLR (Unidirectional Link Routing) and co-chaired by Walid Dabbous,
Head of project RODEO (renamed PLANETE since 2001) and Yongguang Zhang (Hughes
Research Labs). “We needed one year of theoretical discussions before
we could even think about the first developments,” remembers Walid Dabbous,
“Then our team worked for two years on software, computer boards, experiments
and development.” In addition to RODEO members, Japanese researchers from
project WIDE (a group of Japanese scientists and industry representatives) participated
in the design of the standard.
The protocol, also called UDLR, and the deployment of services were tested on
satellite links independently by INRIA and a Japanese team. In April 2001, the
researchers' efforts were crowned with success. After four years of validation
by the international community, their protocol was recognized by the IETF under
RFC 3077 (Request For Comments), an indispensable open sesame prior to any standardization.
The protocol is now validated as a “proposed standard.”
A startup to boot
In June 2000, bolstered by their success, four INRIA researchers-engineers (Emmanuel
Duros, Luc Ottavj, Patrick Cipière and Antoine Clerget) together with
Didier Tymen launched themselves in industry in order to market and integrate
the UDLR protocol into ready for use software. They founded the UDcast company
at Sophia Antipolis.
In the meantime, the advent of broadband, especially that of ADSL over phone
lines, had considerably diminished the interest of satellite Internet integration.
In effect, the ADSL technology is cheaper and offers better performance with
shorter transmission delays, which allows for better interactivity. Nonetheless,
satellites remain an interesting solution to establishing broadband links in
rural areas, or in countries with problematic connectivity even in urban areas.
They also make it possible to secure terrestrial corporate networks against
natural or criminal risks. Emergency satellite links offer the guarantee of
a separate, direct and immune path in case of failure of terrestrial equipment.
Satellites and interactive digital TV
In addition, the UDLR protocol as it was designed is not specifically dedicated
to satellite links, but above all to unidirectional routing. Today, it finds
an increasing number of applications in entirely different contexts. This is
the case for example in terrestrial digital audio streaming or in digital television
for portable devices, in particular to make such broadcasted services fully
interactive. The UDLR protocol also paradoxically regains interest in terrestrial
networks of TV on demand contents distribution over fiber optics, which are
characterized by heavy downlink throughput towards the user and light uplink
throughput. The protocol makes it possible to practically divide the number
of fibers by two and reduce the cost of coupling devices (the interfaces used
to send data into the fiber optics).
The excellent health of UDcast is proof of the multiple applications of UDLR.
The company now has a staff of 22 with an office in Paris and one in Washington
D.C. The company focused on solving performance and security problems concerning
the distribution of IP protocols and offers Internet access and application
providers products, hardware and software. It has established a strong partnership
with Nokia to distribute data or TV on mobile phones using the frequencies of
digital terrestrial TV.
Two INRIA teams are contributing to group distribution, that is to say sending
data to a large number of addressees at the same time.
In certain cases, it is necessary to send the same set of data to millions of
addressees over the Internet. It is possible to do this very economically with
a single operation on the part of the sender by using what is called multicast
distribution. This process is of interest to software publishers to inform their
numerous clients that an update is available, or to inform subscribers that
such and such new service can now be accessed. Similarly, when several persons
wish to participate in a videoconference, the same set of audio or video data
must be sent to a large number of addressees.
In order to cut distribution costs down to a minimum, the network itself must
be capable of duplicating the information and relaying it to the addressees.
The parts of the networks and machines that are not concerned by the distribution
must not be congested by traffic that is of no interest to them. In addition,
the data must arrive without corruption, which is what specialists call transmission
reliability. In the case of one addressee, reliability is ensured by a two-way
exchange: the addressee sends back control messages to the source until complete
reception of the data.
Similarly, for the time being, most reliable multicast Internet protocols work
bidirectionally. However, on a large scale and for a large number of addressees,
this solution is not satisfactory. It entails the management of considerable
quantity of back messages that flood the source and network. In this case, the
solution consists in establishing a unidirectional link, without back route,
and to define specific transmission control mechanisms to ensure reliability.
The solution to reliable transfer of large files
A work group called RMT (Reliable Multicast Transport) was created at the IETF
(Internet Engineering Task Force), the Internet protocol standardization body,
in order to deal with the problem. Members of project PLANETE have been participating
in RMT since 2002.
This work group develops a reliable unidirectional protocol for large scale
data distribution called ALC (Asynchronous Layered Coding) that makes it possible
to avoid data loss problems. The data is sent redundantly until the user terminates
the reception once all the data is received.
INRIA researchers participated in the development of an ALC-based solution adapted
to the specific problem of file transfers. This solution is called FLUTE (File
delivery over unidirectional transport). It explains how to best make use of
ALC in order to provide a file distribution service, in particular for “large”
files. FLUTE defines the mechanisms necessary to transmit files, transport metadata
(information pertaining to the data, such as the name, size, encoding, etc.)
and set up filters on the receiving end.
FLUTE was tested and validated in December 2003 with Nokia and the University
of Tempere in Finland. It was then standardized by the IETF (RFC 3926, Request
For Comments) in October 2004. FLUTE is already a key element in the distribution
of multimedia data over radio networks, in the process of being standardized
at the 3GPP (3rd Generation Partnership Project). The 3GPP is the organization
that proposes standards for GSM, GPRS and UMTS devices, as well as for Internet
distribution on telephones and other portable terminals, in the framework of
the DVB-H (Digital Video Broadcasting, Handheld) standardization. The DVB-H
is an industrial consortium dedicated to standards for digital television on
small mobile screens.
Routing protocols adapted to multicast distribution
Multicast distribution poses specific problems from the point of view of routing,
either to ensure data transmission quality, or to optimize the transmission.
The ARMOR team in Rennes (especially Bernard Cousin) has proposed three multicast
protocols at the IETF since 2001. These protocols are compatible with the IPv4
and IPv6 data transmission Internet protocols. The first protocol makes it possible
to force distributed packets to go through a specific path in the network, thus
avoiding congestion points (something similar to "alternate routes”
to go around traffic jams on the road). The second protocol accelerates the
distribution of data packets to the addressees by diminishing the load of network
equipment. The network equipment located on the path but not in charge of duplicating
packets does not have to process these packets. Lastly, the third protocol is
specifically intended for small groups. This is typical of the case of networked
games where the number of players for each game may be small, but the number
of simultaneous games may be very large.
INRIA researchers were among the first to think up broadband wireless communication
networks between mobile computers. They are still on the forefront today in
the standardization of the protocols necessary for a new generation of wireless
networks, the ad hoc networks.
Let’s go back in time. Before wireless networks, corporate network architectures
were especially complex and fixed. Each computer was connected to the others
via a quantity of wires managed by a powerful network server. In a second stage
around twenty years ago, a revolutionary solution was developed, considerably
lighter and economical—Ethernet. A single coaxial cable connected every
machine to the server. This time, several communications had to share the cable
and computer resources.
A team of a dozen INRIA researchers got interested in the subject of how to
improve wire networks and decentralize the management of communications, in
order to avoid having to go through this central “policeman” that
was the server to go in or out of the network. “We rapidly realized that
the last element penalizing these networks was... the wire,” says Philippe
Jacquet, Head of project HIPERCOM at Rocquencourt. “We started to think
about a high speed wireless Ethernet network, possibly with mobile users. We
felt there could be many applications in offices, vehicles...”
As surprising as it may seem today, at that time Ethernet networks were just
starting to be deployed and most of the industry was very dubious about the
idea of wireless computer networks. Some companies were interested however,
such as Dassault for military applications and Apple for personal computers.
“We continued our work and participated in the development of an original
standard called Hiperlan, defended at the ETSI, the European Telecommunications
Standards Institute. Hiperlan brought two concepts together, one developed for
Ethernet and the other (a self-routing principle) stemming from the Internet
world,” continues Philippe Jacquet.
Hiperlan: a standard ahead of its time
Self-routing between computers was designed for the Internet at the beginning
of the 1990s. It looked like a solution to increase the range of wireless network
communications. Indeed, quite surprisingly, physics commands that the higher
data rate, the shorter range, due to signal scrambling problems caused by the
echoes engendered by such obstacles as walls and furniture.
HIPERCOM researchers started to work on this solution in the framework of a
1994-1996 European project with Dassault, Electronica (I), Symbionics (UK) and
the Universities of Bradford and Bristol (UK). This concept of distributed network
with internal routing was also at the heart of the Hiperlan European standard
(cf. What is a wireless telecommunication standard?) defended by several companies
including Apple, Motorola, ATT, Canon and Daimler-Benz. Hiperlan was however
probably too ambitious. It aimed at transfer rates of 25 Mbits/s. “It
was kind of the Concorde of the wireless networks,” remembers Philippe
Jacquet with nostalgia. At the time, no company dared develop a product.
Wireless networks, the infamous Wi-Fi, were finally deployed in 1996 according
to a competing standard called 802.11, defended at the IEEE by some of the same
companies that were working on Hiperlan. The IEEE (Institute of Electrical and
Electronics Engineers) edicts international standards, in particular for telecommunications.
When Europeans were trying to develop a totally new standard and products, the
Americans modified their products, and were able this time to impose industrial
solutions more rapidly.
Nonetheless, the competition with Hiperlan certainly influenced the Wi-Fi standard
significantly, in particular concerning the type of network created, easy user
access and routing mode. Wi-Fi networks are thus designed around a single access
point with which several computers can communicate using a very simple distributed
protocol, just as Hiperlan proposed, that makes it possible for users to freely
enter and leave the network.
Another consequence inherited from Hiperlan concerns the ability of the network
to function in ad hoc mode—a mobile network in which the various computers
connected at a given moment can serve as relays between one another and automatically
route data. This type of functionality had been developed in Hiperlan to alleviate
the short range problem due to the high rates envisioned. In the same way, ad
hoc networks can be used to extend the range of Wi-Fi networks or to design
a wireless network with an off-center access point, as is often the case. Thus,
mobiles that are out of range of the access point can use other mobiles as relays
to forward their communication to the access point.
Concepts that are now applied to the new wireless networks
In parallel with the development of these radio technologies, the Internet protocols
also had to be adapted to computer physical mobility that causes frequent changes
in network topology. The interoperability of the various networks had to be
ensured as well as their stability in order to avoid interruptions in transmissions,
without eating up all the available bandwidth with the traffic necessary to
describe this topology. In 1996, the Internet protocol international standardization
body, the IETF (Internet Engineering Task Force), launched a work group on the
topic called MANET (Mobile Ad hoc Network). “They were interested in the
concepts we had developed in Hiperlan, so they quickly got in touch with us,”
remembers Philippe Jacquet. “They wanted to propose our OLSR (Optimized
Link State Routing) routing protocol, adapted to the Internet and optimized
for mobile networks. OLSR was the first protocol that made it possible to experiment
on plug and play ad hoc networks.” The protocol was an instant success
and was downloaded more than 2,500 times! For the time being, ad hoc networks
are primarily of interest to the military, automobile manufacturers and Wi-Fi
enthusiasts. The latter are carrying experiments between several buildings in
Berlin and in Paris, using INRIA's protocol.
Today, about fifteen researchers keep working on OLSR. The protocol is experimental
(RFC 3626, Request For Comments), as are the three other competing protocols
at the IETF, all of which are American. “The breakthrough of OLSR at the
IETF in 2000 owes a lot to Thomas Clausen, a Danish doctoral candidate who joined
us in 2000 and who knew how to communicate adequately in such forums.”
OSLR is probably the most implemented protocol presently, with around fifty
implementations in France, the United States, Japan, Canada, Norway and Finland,
among others, including by several companies such as Boeing and Cisco. Cisco
is planning to market hybrid OSLR-based servers within a few months that will
be able to operate with or without wires. The ideas developed in OLSR have also
been reused in other IETF works, and several standard proposals by large American
corporations include OSLR concepts.
Planning for the integration of third generation devices
In addition to all of the above, wireless networks must face new needs, such
as transmitting multimedia data to cell phones or PDAs, the so-called 3rd generation
devices. The companies and research departments that develop such technology
participate in the standardization of such networks within the 3GPP (3rd Generation
Partnership Project). The 3GPP was created in 1999 and regroups several organizations
such as the ETSI. It defines the standards adapted to mobile telephony: GSM,
GPRS and UMTS.
One part of the ARMOR team at Rennes is working on the mechanisms required to
manage the displacements of mobile terminals in such a way that the latter may
be able to transparently use several types of radio networks (UMTS, Wi-Fi, Wi-Max,
and so on). In this way, it will become possible to use multimedia applications—voice
over IP, videoconference, video on demand, etc.—through the best communication
means available at every instant and every place. The different technologies
thus cooperate with each other in order to supply the best possible covering.
Researchers are also developing data compression techniques for efficient transmission
of multimedia information on GPRS and UMTS networks. More precisely, they are
proposing an IPv6 header compression method at the 3GPP. The method is already
standardized at the IETF under the name of ROHC (Robust Header Compression).
By definition, wireless communications imply radio wave transmissions. The radio
resource is very much in demand by radio stations, TV channels, satellite links
and the military. To develop wireless telecommunication equipment, it is thus
compulsory to go through standardization in order to obtain a specific frequency
band. The ETSI is in charge of this in Europe, in agreement with the various
national jurisdictions, such as the AFNOR, the French standardization body.
Concretely, each standard (GSM, Wi-Fi, Bluetooth, OLSR)must specify the power,
modulation mode and protocols to be implemented in dedicated equipment in order
to use this frequency and interoperate without interfering with other devices.
Several INRIA research projects are developing languages to improve the design
and processing of Web documents, not only in terms of structure, but also in
terms of meaning, in order to create what is known as the semantic Web. Such
work is carried out in the framework of the W3C international consortium.
When the Web got started, the problem of page representation rapidly arose.
A standard on the subject called SGML had been published by the ISO (International
Standardization Organisation) in 1986. This was a first step toward a structured
approach to documents, a logical, explicit organization into chapters, sections,
subsections, and so on. Such an organization makes it possible to process the
document's contents, for example to search by keyword, and to more easily navigate
it.
Researchers of the WAM team (formerly known as OPERA) of INRIA Rhône Alpes
had been working on such logical document representations for about ten years.
They thus naturally got involved in the work concerning the Web page representation
format. In order to experiment on these formats, they had developed a prototype
software. “When the W3C (World Wide Web Consortium, in charge of developing
Web standards) was set up in 1994,” recalls Vincent Quint, Head of project
WAM, “it got immediately interested in our software. The W3C needed such
a tool to validate, experiment and demonstrate new Web technology, especially
from the point of view of the users who either produce or make use of documents.
Still up to date software
Two team members thus joined the W3C technical team (see box “INRIA, a
major W3C player”) to pursue the development of the software, Amaya. Amaya
thus progressively integrated the new languages created by the W3C. As time
went by, it became an operational authoring tool that not only makes it possible
to use the most recent Web technologies, but also to produce complex Web pages
containing text, graphics and mathematical expressions, that conform to W3C
specifications. A brand new version of this Web editor was released at the end
of 2004.
INRIA also widely contributed to developing some of the new formalisms, such
as MathML, which is the W3C standard to represent mathematics in XML documents.
XML is the new Web page representation format, the successor of HTML. MathML
makes it possible for teachers, students, researchers and engineers to put math
on their Web pages and to exchange it by email or from one software to another.
MathML is the result of a work group created in 1997 by the W3C. Researchers
from project CAFE (formerly SAFIR of Sophia Antipolis) participated in it from
the start. In the 1990s, they had already developed a standard in this field
called OpenMath based on SGML, in collaboration with other research institutions.
Actually, MathML makes it possible to use OpenMath to describe mathematical
objects that are more complex than those natively represented in MathML.
The first MathML recommendation was issued in 1999. The second version dates
from 2003. A certain number of documents are already using MathML, including
the American patents of the US Patents and Trademarks Office. Large scientific
publishing houses such as Elsevier and Springer, as well as online education
publishers, are expressing interest and will use it as soon as they start encoding
their documents in XML.
The multimedia puzzle
Another concern rapidly assumed considerable importance on the Web—the
development of multimedia, either static as with images and text, or dynamic
with sound and video. Such documents had to be guaranteed to remain usable in
such heterogeneous environments as a computer, a cell phone and a TV set. Several
W3C work groups were set up to devise solutions. Nabil Layaïda, a project
WAM researcher, has been working for several years in one of these groups called
“Synchronous multimedia”. The group was created at the beginning
of 1997 with the goal of adapting multimedia documents to the Web, defining
temporal relations between the different information elements in a document
and planning how sound, images and video will fit together within the space
of the screen as well as in time. This work group develops a language called
SMIL, to which INRIA researchers greatly contributed, especially through concepts
developed by Nabil Layaïda during his doctoral thesis. The first SMIL version
was standardized in June 1998, the second one in August 2001. Version 2.1 is
done since may 2005. The format is already widely used, for example by Realplayer
and by MMS multimedia messages that succeed SMS messages in a version adapted
to cell phones.
Nabil Layaïda also coordinated the development of software implementing
SMIL. One of these Web tools called Limsee has been available since the summer
of 2004. It makes it possible to create adaptable multimedia presentations in
the SMIL format. Another one called PocketSMIL is dedicated to PDAs and portable
devices.
In the same spirit, the W3C created another work group called “Device
independence” the goal of which is to make sure that the Web remains independent
from the devices used to access it. A doctoral candidate of project WAM, Tayeb
Lemlouma, has participated in this group until 2004. The research concern for
example the transformation and adaptation of a multimedia document including
video, sound and text to a cell phone. The solutions consist in replacing the
video by still images, or to restructure the documents to display them sequentially
When computers start reasoning...
Nonetheless, beyond such needs for document structuring and data processing,
we still must face the barrage of information coming from the Web. One of the
solutions prepared by the W3C since the end of the 1990s, intends to make document
contents more intelligible, to give meaning to the information stored on Web
pages in HTML. This is what is called the semantic Web. XML standardization
then is a first step: it defines the document and data structure syntax. To
access the meaning, semantic Web languages then make it possible to organize
and prioritize the concepts used to describe Web resources into ontologies.
Ontologies are logical structures that capture a certain number of logical relations
between the concepts. Such languages are for example capable of deducing from
the fact that “you need a ticket to take a train” and the fact that
“the TGV is a train”, that then “you need a ticket to take
the TGV”. Ontologies organize the description of concepts, such as “ticket”
or “train”. From then on, information search can be carried out
intelligently by the computer itself, automatically and without user intervention,
simultaneously on several sites and independently of data format. For example,
a computer can plan a trip for a given destination involving planes, trains
and hotels. In fact, the semantic Web makes it possible for computers to reason,
associate neighboring concepts together for a given request, all things that
are impossible with today's search engines. The answers will be more precise,
more relevant and the information retrieved will be correct.
The first semantic Web language called RDF (Resource Description Framework)
was standardized by the W3C in 1999 and 2004, followed by RDF Schema standardized
in 2004. RDF allows simple semantic descriptions, and RDF Schema supplies a
basic vocabulary to describe the meaning of the concepts used. Two INRIA projects
are particularly involved in semantic Web work, EXMO in Grenoble and ACACIA
at Sophia Antipolis.
Since 2001, EXMO researchers who had been working on the design of knowledge
representation languages, naturally contributed to the “WebOnt”
W3C work group dedicated to developing a third semantic language that is more
expressive than RDF Schema. This language is called OWL and was standardized
in February 2004. It is the first language capable of defining ontologies. Several
software packages using OWL are under development; Operational systems are being
produced by Hewlett Packard and the Universities of Manchester and Karlsruhe.
From the first applications to tomorrow's search engines
These languages, especially RDF and RDF Schema, are beginning to be used. The
main application in the world is called FOAF (Friend Of A Friend). It was created
to connect people, create acquaintance networks and partnerships. Everyone describes
his or her profile (name, email, interest, profession, etc.) and the computer
does the rest. The semantic Web is also of interest to companies to manage their
knowledge.
To ensure the development of RDF and OWL, the W3C launched a work group to define
the “Semantic Web Best Practices” in 2004. The goal of the group
is to define the languages, to offer methodological elements, answer utilization
questions and provide pedagogical material. Fabien Gandon of project ACACIA
participates in the work group. Project ACACIA is interested in methods and
tools for knowledge management. Since their goal is to ensure the interoperability
between different solutions, their work is also in the context of the semantic
Web. A platform called CORESE has been developed since 1999 that makes it possible
to design servers dedicated to the semantic Web. These servers are based on
a search engine that exploits descriptions of the semantic contents of documents.
CORESE implements a translator to read and produce RDF descriptions by interpreting
them in the conceptual graph formalism, a method to represent knowledge and
reasoning that benefits from twenty years of research. The CORESE platform is
available on the Web.
Finally, for the semantic Websemantic Web to really be operational, especially for search
engines, request languages for RDF and OWL must also be designed, in order for
example to simultaneously exploit two different ontologies and make sure they
are interoperable. A W3C work group called “RDF Data Access Group”
is devoted to the problem. Researchers from project EXMO belong to this group.
It is in this spirit that Olivier Corby of project ACACIA is evaluating the
performance of the request language he designed for CORESE.
INRIA has been one of the pillars of the W3C (World wide web consortium), an
international consortium that ensures the development and promotion of Web standards.
The institute was the first European host site, from 1995 to 2002, along with
MIT (Massachusetts institute of technology) for the American continent and the
University of Keio in Japan for Asia. Since 2003, ERCIM (European Research Group
for Computer Science and Mathematics) has taken the relay from INRIA in Europe.
During these 8 years, 20 people from INRIA participated in the W3C technical
team (which counted some 60 members in all). Jean-François Abramatic
moreover presided over the consortium for 4 years until 2001. Vincent Quint
was responsible for one of the four W3C technical fields, that is, the format
of documents used on the Web and user interfaces. "This was above all a
guiding role, he explained. This involves keeping an open mind to needs, coordinating
efforts, suggesting the creation of working groups and ensuring the participation
of researchers and industrial companies."
Currently, 6 or 7 INRIA researchers participate in working groups. Vincent
Quint, director of research for INRIA, has been co-chair of the "Technical
Architecture Group" (TAG) for W3C since February 1, 2005.
Following is how the LORIA (Lorraine Research Laboratory in Computer Science
and its Applications), a joint department of INRIA, CNRS and the three Nancy
universities, became a national and now international reference site in the
field natural language computer processing, and especially in standardization,
which is the work topic of a dozen researchers.
Linguistic computing was invented practically at the same time as computer science.
Right from the start, the idea appeared of using computers for automatic translation,
a feat however still out of reach today. The deployment of the Internet and
the multiplication of electronic documents then only increased opportunities
and needs in indexing, classification, text search and dialog transcription
in any language and independently of their evolution over time.
The whole difficulty is to achieve generic solutions that are applicable on
an international scale, and that can be specifically parametrized for a given
language or a special need. In order to get an idea of the complexity of this
task, it is enough to take the example of a lexicon (a dictionary) and the notion
of adjective with all possible declension forms: in French, you will have the
“masculine” and “feminine” genders and the “singular”
and “plural” numbers, whereas in Japanese you will also have to
consider the possibility of negation and tense agreement. It thus appeared clearly
since the beginning of the 1990s that the only solution to achieve a perpetual
management of the world's linguistic resources was to go through standardization.
A way of speaking the same language
The first international standard in the field concerned terminology, that is
to say the vocabulary that is specific to this and that industry, science or
institution. The ISO, the International Organization for Standardization, has
been thinking about this problem since its inception in 1947, within a dedicated
technical committee (TC 37). Indeed, by definition, any standard uses a specific
terminology and its multiple translations. It is also easy to understand the
interest of such an approach in view of the 360.000 pages of European Community
institutions texts that now have to be translated into 25 languages... a titanic
task.
“The ISO called on us in 2000 on the basis of our work in linguistic modeling,
in order to try and find a standardization solution in terminology,” says
Laurent Romary, Head of project Langue et Dialogue at LORIA. “At the time,
two standards were in competition, an American and a European one. Using our
more abstract approach, we were able to unify the two within a common specification
platform, which is now a reference.”Laurent Romary is the editor of this
standard (ISO 16642, or TMF for Terminological Markup Framework) that was published
in 2003, only three years after his first dealings with the ISO.
The interest of a standardized platform is that it makes it possible for a given
user, say a company, to create their own terminological format, specific to
their activity, and to exchange documents on this basis with for example subcontractors,
service providers or clients who adopt the same format—in a sense it provides
a way of speaking the same language. The ISO 16642 standard was very successful
and has already been implemented many times, including by IBM and by Daimler-Benz.
A white paper published in 2005 describes how to use it with concrete application
recommendations. This work will result from a collaboration between LORIA, such
industry partners as EDF and EADS and institutional partners such as the INIST
(Institute for Scientific and Technical Information, the CNRS information center),
in the framework of a national INRIA research and development initiative called
SYNTAX
Dealing with large volumes of electronic documents
Another international organization took an interest in INRIA work, the TEI (Text
Encoding Initiative) consortium. The TEI is a gathering of international institutional
partners interested in managing large quantities of electronic archives. The
consortium was created in 1987 for purposes of defining perpetual text formats
for libraries, universities, museums, publishers and so on. Given the quality
of the work carried out by the consortium, INRIA first took inspiration from
its representation directives to publish its own written documents or dialog
transcriptions. “We progressively integrated our own tools to annotate
the texts and prioritize the information, etc.” says Laurent Romary. “Our
work then attracted interest on the part of the consortium and in 2000 we were
asked to participate in its Scientific Board, which we still do today.”
What is even more rewarding is that the LORIA and two CNRS units, the INIST
and the ATILF (Computer Analysis and Processing of the French Language), constitute
one of the four TEI host sites, next to the University of Virginia and the University
of Providence in the United States, and Oxford University in Great Britain.
Nancy research scientists will bring their skills in data modeling to define
more generic text formats. On the national scale, this collaboration also makes
it possible to deploy TEI standards in specific contexts, for example to standardize
gray literature (scientific production, activity reports, doctoral dissertations,
and so on).
Standardizing lexicons and contents
There is thus nothing surprising for a LORIA researcher, Laurent Romary, to
be called on to chair the new ISO subcommittee dedicated to the standardization
of linguistic resources that was founded in 2002. Its objective is to standardize
all the information necessary for linguistic engineering, for example for orthographic
or grammatical correcting, automatic translation, information extraction, etc.
In this framework, a team associating INRIA (Gil Francopoulo) and the U.S. Department
of Defense is developing a standardized platform called LMF (Lexical Markup
Framework), this time to represent wide spectrum lexicons rather than terminologies.
They plan to co-publish a standard that models the representations associated
with words. The work involves in France a national network of more than fifty
industry and institutional contributors.
The subcommittee is also working on another standard format to represent contents
called MLIF (Multilingual Information Framework). The standard would be adapted
for example to translation memories (typical phrases created by translators
when frequently occurring), the localization of certain key messages in software,
DVD subtitling, among others.
An INRIA research team (MERLIN) has been participating for many years in standardization
activities in software ergonomics.
The numerous benefits of computers and more generally of information and communications
sciences must not overshadow the fact the primarily concerned party are... the
users. A computer system, as significant and advanced as it may be, offers little
interest if it does not obey certain criteria of usefulness, health and security,
but also comfort and user-friendliness, which all are usability characteristics
that must be taken into account from the design stage.
Computer technology has now invaded the private and professional spheres—work,
transportation, services, leisure. They are part of all aspects of life, especially
in developed countries. It is thus not surprising that the ergonomics of such
systems is becoming a growing concern. This means better understanding human-machine
interaction in order to improve user comfort and security as well as system
efficiency.
Computer ergonomics has been taken into account by the ISO, the international
standardization organization, and its national authorities such as the Afnor
in France (French Standardization Association), since 1980. The software aspects
(design process, dialog techniques, multimedia,...) are concerned, as well as
the ergonomics of the work station, its surroundings and of the hardware. Computer
ergonomics problems are even increasing. The new usages and issues of information
technology require adaptation. In fact, more and more experts are appointed
to participate in the ISO international meetings. Governments, institutions
and software publishers are increasingly concerned.
“The topics we are working on now concern the Web and graphical interfaces,
virtual reality, for example,” explains Dominique Scapin, Scientific
Head of project MERLIN (INRIA Rocquencourt and Lorraine) that has been contributing
to software ergonomics standardization since 1988. In France, twenty-five standards
and specifications concern computing ergonomics, and fifteen are more specific
to software ergonomics. MERLIN researchers published many articles and manuals
for software designers. Recently, Dominique Scapin who organizes the “Software
Ergonomics” group of the AFNOR, coordinated the publication of the first
AFNOR compendium of standards, specifically on computer ergonomics.
New usages, new populations
In practice, what is meant by software ergonomics? In fact, there are many facets
to this issue. For example, design processes are concerned in order to help
take into account user characteristics (correctly taking into account the usage
context and user experience and knowledge). The best practices for dialog between
user and software interface are also defined (degree of adaptation, conformity
to user expectations, easy learning, type of interactive menus, etc.). Rules
to present information and guide users right are proposed. Methods to evaluation
software usability are also defined. All these topics are at the heart of project
MERLIN research work, with the goal of improving the ergonomic quality of interactive
software. On the one hand, the team works on integrating ergonomics results
into new software design, and on the other hand, it is interested in new computing
applications such as multimedia and new populations of users.
This preoccupation actually concerns a large number of research works in the
world, especially about the concept of accessibility, that is to say adapting
computer systems to the largest number of individuals, irrespective of their
capabilities, or sensory, motor or cognitive impairments (see box). Among other
things, this includes dealing with an aging population. A law on this subject
was voted in France in July 2004 by the Senate. Several laws were also adopted
in the United States and European directives are in preparation. Four standards
are being finalized in Japan. Consistently with these reflections and upcoming
laws, the ISO is working on a technical specification on the subject. A first
standard (ISO/TS 16071) was published in 2003. Its goal is to guide developers
in human-machines interface design that propose the highest possible level of
accessibility.
" Our team is working in all these areas ", says Dominique Scapin,
“especially in the new areas of knowledge (for example, understanding
the problems raised by Web software or virtual reality software) and on interactive
software design and evaluation methodology.” Even if it is difficult to
ascertain the importance of such and such contribution, this research has a
recognized impact on the publication of standards in this field.
Accessibility is a concept that encompasses the differences in capabilities dues to age, illness or handicap. It concerns persons who suffer from physical, sensory or cognitive deficiencies from birth or acquired during their life, elderly persons who may benefit from new products and services but who have reduced physical, sensory and cognitive capabilities, persons with a temporary impairment, for example someone with a broken arm or people who have forgotten their glasses, and persons faced with difficulties in certain situations, for example a person working in a noisy environment or whose two hands are occupied by another task.
In the last ten years, free (de facto standards) or standardized software technology
has changed significantly. Such approaches as component models and model engineering
now are the core of much research, especially at INRIA.
Software design and everything that goes with it (development tools, assembly,
tests, deployment, execution, administration, and so on), which is globally
referred to as software engineering, is changing following the increasing complexity
of developed systems. Such systems are more and more often embedded, distributed
over several machines and work in real time. It is now commonplace for a company
to develop software with several million lines of code. On the other hand, productivity,
quality and flexibility demands have continuously risen. All these changes entail
a definite need for automation.
The main current developments in software engineering partly stem from the deployment
of object technologies, which are now widespread in computer system analysis,
design and implementation. In 1991, the OMG (Object Management Group), the standardization
organization dedicated to such technology, defined a reference software architecture
for distributed applications—several components that cooperate over several
sites—using object technology. This architecture called Corba makes it
possible for applications developed in different languages to communicate with
each other, even if they are not on the same computer. Corba is a kind of middleware,
a class of software that takes charge of intermediate functions between data
transport and the applications, such as communication, task allocation functions
and such properties as security. This standard is widely used in industry, especially
for corporate server information systems and in command and control military
applications, as well as for certain applications in industry and telecommunications.
Automatically deploying software for distributed applications
Being interested in the design of complex distributed applications with specific
constraints, research scientists of project JACQUARD (INRIA Futurs) rapidly
took an interest in Corba. In order to solve certain problems in the use of
the software, they developed a scripting language for Corba (a programming language
with specific properties) in the context of Philippe Merle's doctoral thesis.
Thus, in 1997 when the OMG issued a call for proposals on the topic, the team
proposed its solution called Corba Scripting Language, which was accepted in
2000. Researchers got familiar with the functioning of the OMG on this occasion
and followed the work carried out in the other work groups, concerning in particular
models based on software components, an approach initiated by the OMG in 1997
that is especially adapted to distributed applications. As a matter of fact,
object models make it possible to build software but cannot for example automatically
deploy software over a set of machines, as is necessary in distributed applications.
Component based models transparently support this functionality. JACQUARD researchers
soon joined the dedicated work group and finally took the helm thereof in 2000.
The Corba standard integrating the component model was issued in 2003.
Up to 2002 they were simultaneously experimenting on component models in their
own prototype software, a platform called OpenCCM developed in the framework
of the ObjectWeb consortium —a consortium created by INRIA, France Telecom
and Bull to develop freely available, quality middleware. Research and development
was then continued in the framework of national (RNTL IMPACT, RNRT COMPiTV, ACI GRID RMI) and European projects (IST COACH, ITEA OSMOSE), especially with
Thales and a Greek company called Intracom. OpenCCM has been freely available
over the Internet since 2003. It was downloaded more than 2,500 times and is
one of the two OMG reference platforms to implement the OMG component model
standard.
At the beginning of the 1990s, no less than about fifty object modeling methods
had been developed. At the end of 1997, in order to maintain a certain uniformity,
the OMG standardized the first unified object modeling formalism, UML (Unified
Modeling Language). Industry rapidly adopted this code generation technique
from models, an approach based on graphical representations. Each company developed
its UML variant, dedicated to its applications. These are called dedicated languages
and UML has become one dedicated language among others. Currently, the two main
players in the field are IBM and Microsoft. Both have developed their own implementation
platform and associated tools: Eclipse for IBM and Visual Studio for Microsoft.
The ambitious concepts of model engineering
One of the solutions that could make it possible to harmonize such slightly
anarchic developments and ensure the compatibility and interoperability of these
dedicated languages is model engineering. This software design methodology is
the object of much research throughout the world. It was accepted by the OMG
in November 2000 as a major approach called MDA™ (Model Driven Architecture).
The objectives are multiple and ambitious: to automatically generate code, to
easily and automatically pass from one model to another, from an abstraction
level to another, or from a given application to another in order to automatically
generate software component tests, for example. The approach is based on several
already established technical standards: such languages as UML, modeling languages
such as MOF (Meta Object Facility) and above all model transformation languages
such as QVT (Query View Transformation), which is in the process of being standardized.
QVT is a model transformation language specification that makes it possible
to adapt a model to new constraints, and to transform any dedicated language
into another. This language lies at the heart of the success of the MDA approach.
The OMG is planning to standardize QVT in december 2005.
Two INRIA teams, TRISKELL in Rennes and ATLAS in Nantes, are carrying out research
in model engineering. They explored two different and complementary directions
concerning model transformation to adapt it to the users, and developed two
model transformation languages, which are in a way QVT implementations: ATL
(ATLAS Transformation Language) by ATLAS and MTL (Model Transformation Language)
by TRISKELL.
ATL is a model transformation language intended for data engineering, in other
words for information systems and database management systems. In effect, the
size of such systems keeps growing, the data they contain is increasingly heterogeneous
and complex, and the applications making use of them more and more varied. ATL
stems from research carried out in collaboration by INRIA, the University of
Nantes and company located in Brest, TNI-Software. ATL and its programming environment
on IBM's Eclipse platform have been officially available as free software since
October 2004, on the Generative Model Transformer site. ATL was awarded the
IBM Eclipse Innovation Prize that same year. Simultaneously, the team is implementing
the same principles following Microsoft's approach, in the framework of a research
agreement, and is extending the functionalities of the operators beyond transformation
by taking into account model weaving for example. This work is mainly done in
industrial collaboration projects, such as the ModelWare European project steered
by Thales. Close partnerships with other French companies, such as the Sodifrance
group, made it possible to develop other tools that complement ATL in the framework
of a research platform in model engineering (AMMA).
MTL is a generic object language, a basic building block for developing frameworks
(a kind of application sketch) of model transformation. TRISKELL is interested
in frameworks allowing for powerful semantic operations based on algebraic operators.
MTL is for example especially well adapted to the control of software components,
with the goal of ensuring the reliability of their assembly in the context of
distributed applications. The language makes it possible to automatically generate
software component tests. It is another way of taking advantage of model engineering,
by changing representation mode. As a matter of fact, testing software components
and verifying their reliability before assembly has become a sensitive and often
very costly step before the deployment of an application. TRISKELL researchers
had already worked on the topic in the framework of a national project with
France Telecom, Gemplus and Softeam to define test techniques in UML. MTL is
available online as free software (modelware.inria.fr).
Proposing solutions for different application areas
In addition, distributed applications are more and more embedded and in real
time as in cell phones or car and aircraft systems. To model such complex embedded
systems, AOSTE researchers (Sophia Antipolis) have been developing for twenty
years so-called reactive synchronous formalisms, called RTE-ML (Real Time Embedded-Meta
Language) such as Esterel and the SyncCharts. In the final analysis, such languages
are rather close to UML state and activity behavioral diagrams that are indispensable
in the context of distributed applications. The approach was of immediate interest
to the companies that were developing UML for embedded systems, led by Thales.
In 2001, AOSTE researchers designed an UML-compatible RTE-ML profile in the
framework of a European project with Thales, Esterel Technologies (an INRIA
startup), Nokia and Finnish universities.
ATLAS researchers are also interested in a particular application of model engineering:
finding generic solutions to automatically renew the old computer systems programmed
several dozens of years ago (for example in Cobol, PL/1 or Ada) in order to
adapt them to modern technological environments. The OMG launched an initiative
of the subject in 2003 to solve this large scale industrial problem, the importance
of which was already highlighted on the occasion of the year 2000 and that of
switching to the euro. The AMMA model engineering experimental platform with
its various components (transformation, weaving, projection, global management
of models) is the basis of such reverse-engineering for information systems
which often entails hard data engineering problems. Certain tools developed
by ATLAS should be directly applicable to this especially important industrial
problem.Carroll: INRIA, a partner that cannot be ignored in software engineering
In March 2003, in the framework of the development of their electronic
defense systems (sea, land and air), Thales initiated a three year partnership
called Carroll with the CEA and INRIA in the fields of model software
engineering and middleware technology. INRIA naturally contributes to
this partnership due to its recognized competence in the fields of embedded,
real time computing, distributed applications and middleware. INRIA projects
ATLAS, TRISKELL and AOSTE initiated three of the four Carroll projects.
In particular, ATLAS and TRISKELL are contributing to the development
of a programming language compatible with QVT. In addition, TRISKELL develops
software component test automatic generation methods, and AOSTE researchers
aim at standardizing a solution stemming from their RTE-ML profile at
the OMG, that is adapted to embedded, real time systems. They proposed
a Request for Proposals (RFP) on the needs in the field voted on in January
2005. They should be in a favorable position to answer this RFP based
on their previous work.
In March 2003, in the framework of the development of their electronic defense systems (sea, land and air), Thales initiated a three year partnership called Carroll with the CEA and INRIA in the fields of model software engineering and middleware technology. INRIA naturally contributes to this partnership due to its recognized competence in the fields of embedded, real time computing, distributed applications and middleware. INRIA projects ATLAS, TRISKELL and AOSTE initiated three of the four Carroll projects. In particular, ATLAS and TRISKELL are contributing to the development of a programming language compatible with QVT. In addition, TRISKELL develops software component test automatic generation methods, and AOSTE researchers aim at standardizing a solution stemming from their RTE-ML profile at the OMG, that is adapted to embedded, real time systems. They proposed a Request for Proposals (RFP) on the needs in the field voted on in January 2005. They should be in a favorable position to answer this RFP based on their previous work.
It is easy to forget but numerical computation is one of the basic building
blocks of computing. The precision of the operators and programs, their
reliability and their speed are at the heart of a large number of more
or less critical industrial applications.
It is enough to remember a few of the most famous computer errors of the
past years, such as the Pentium bug in 1994, or the explosion of the first
Ariane 5 rocket in 1996, to realize how reliability and precision are
crucial in machine computations. The primary culprits of these infamous
failures are microprocessors and the so-called numerical algorithms that
use floating point arithmetic. Floating point arithmetic is the format
implemented in computers to approximately represent real numbers (see
box).
At the end of the 1970s, engineers and researchers were able to impose
a uniform standardized solution, in spite of divergent commercial interests.
At that time, each manufacturer (Digital, Cray, Hewlett Packard,...) had
their own floating point number representation, but certain choices were
only meant to speed up computations, maybe with less transistors, even
if that meant producing notoriously false results or go contrary to programming
usages. Even the simplest arithmetic operations led to variable results
from one machine to the next. When personal computers appeared, the idea
of standardizing floating point arithmetic had gained ground. A newcomer
in the field, Intel, wanted to carve out a place on a market in which
it had no legitimacy and chose to rely on the work of an influential scientist,
William Kahan, Professor at the University of California at Berkeley.
The rest is history. Kahan's ideas strongly influenced the topic and he
then was awarded the Turing Award for his contributions. Almost all of
his recommendations have been accepted by the IEEE
(Institute of Electrical and Electronic Engineers), the learned society
that is an authority in the fields of computer engineering. The IEEE 754
standard was thus adopted in 1985 by the ANSI (American National Standards
Institute). The standard precisely defines the correct round-off for the
four arithmetical operations (addition, subtraction, multiplication and
division), and for the square root. Practically all current computers
obey this standard, even if certain functionalities are only implemented
via libraries during execution or compilation.
The best way to be at the heart of a topic
Project ARENAIRE
of INRIA Rhône Alpes has been working in this field since its inception
in 1999, and completely integrates the standardization approach. Its researchers
are naturally involved in the ongoing discussions concerning the revision
of the IEEE 754 standard. “Our technical ideas and our research
work may thus be taken into account in the future standard,” explains
Marc Daumas, a CNRS researchers working in project ARENAIRE. “However,
for us, to participate is first of all a way of exchanging ideas with
the best specialists worldwide, to establish links that will then have
scientific consequences, to see where the wind blows, which all are plusses
to advisedly develop our ideas.” As a matter of fact, the functioning
of this IEEE technical committee is
especially pragmatic. The standard revision committee of about twenty
persons meets every month in California, and for more efficiency, a subcommittee
works weekly to prepare the proposals to be discussed. ARENAIRE members
visit two or three times a year and then pursue discussions remotely.
In addition to INRIA, only a few British and Israeli scientists participate
in a limited fashion in these meetings, which are open and free of access.
“In this field as in many others, there is no point in developing
cool techniques in your own little corner,” says Marc Daumas. “You
have to take into account the experience of scientists, but also the point
of view of industry, and interferences with the work carried out by other
standardization committees.” Hence the interest of being at the
heart of these expert exchanges, which are at the same time a source of
inspiration for research, and a guarantee of quality for the work. It
is however difficult to isolate direct scientific consequences. One thing
is certain, project ARENAIRE is reputed in its field. ARENAIRE researchers
worked for example with Aerospatiale on the choice of floating point use
for the Airbus A380. Other researchers from the project developed complex
algorithms to reduce division computing times for STMicroelectronics.
Since 2002, they have been developing a floating point library for another
ST processor.
Multiple positive consequences
The consequences are sometimes surprising. Thus, following a consensus
reached between several members of the technical committee in 2004, it
was decided to start working on integrating arbitrary precision floating
point arithmetic into the future standard. This is a field in which another
INRIA team, SPACES of Nancy, has unique know-how. This multiprecision
mode of computation is more precise than floating point computation, but
also slower. The SPACES team developed software called
MPFR
(MultiPrecision Floating-point arithmetic with exact Rounding) that could be compatible
with the IEEE standard when it is completed, for a very reasonable overhead
in terms of computing time. The experience acquired by SPACES clearly
sets it ahead of the competition, including the experts of the standard
revision committee.
Another key area to improve computation precision is the certification
of programs and algorithms with automatic tools based on formal proofs.
Such techniques are beginning to impose themselves in cases where human
lives might be threatened by malfunctioning. They also are increasingly
interesting for industry because of their growing power and automation.
Due to ARENAIRE developments, such techniques can now be applied to programs
that resort to numerical computations. The objective of this approach
is to validate the programs and make sure that their implementation is
correct. In 2000, INRIA launched a cooperative research initiative on
the subject called AOC (Certified Computer Arithmetic) involving several
projects (LEMME, ARENAIRE and SPACES). INRIA has long been developing
a proof assistant software called Coq in the context of formal methods,
that checks the detail of proofs. Since 2002, ARENAIRE researchers have
been working on these topics in collaboration with Nasa researchers of
the Langley Research Center and of the National Institute of Aerospace
(Virginia). Historically, Nasa is strongly supportive of another formal
proof assistant called PVS, which is widely used in the United States
and in industry. Floating point arithmetic is used in a project to automate
small airport air control. Due to a collaboration funded by CNRS and Nasa,
there will soon be a joint Coq and PVS specification, which will ipso
facto become a primary world standard. Even if this specification has
little chance of being accepted by the IEEE, because it is expressed in
a formalism that most standard revision committee experts do not master,
its noted contributions once more make the Institute work better known.
To represent real numbers, which are not integers, most computer systems use
a so-called floating point format. This format has been used for decades in
scientific computing. It consists in defining a real number with a triple: its
sign, mantissa and exponent. Changing the latter causes the decimal point to
“float”. This representation is different from the fixed point representation
in which the exponent is fixed.
Thus, in base 10, the electron mass is 9.109381 x 10-31 kg, that is to say a
plus sign, a mantissa of 9.109381 and an exponent of -31.
On August 9 2005, INRIA (host of the Scilab consortium), Maplesoft, Mathsoft
and National Instruments (Nasdaq : NATI) announced the creation of the
Numerical Mathematics Consortium (NMC). Associated with participants from the
industrial and academic world, these mathematical software publishers joined
forces to establish consistent and concrete bases for numerical programming.
The initial goal of the consortium was the establishment of an open standard
for the semantics of mathematical functions used to develop algorithms for use
in a wide range of disciplines and in various hardware and software environments.
« There has long been a need in our industry for a unified and standardised
base, explains Ali Maleki, Director of the Brake and Chassis Electronics programme
for ArvinMeritor. Today, each tool has its own function set, often requiring
extensive training, which has led us to develop algorithms and know-how that
is not easily transferable to the rest of the industry. We must rewrite these
algorithms for new projects or when implementing new technologies, which leads
to additional costs. A standard set of mathematical functions based on semantics
accepted by the industry would be a great leap forward in the creation of transferable
techniques and ready-to-use libraries and tools that can be used instantly in
a number of environments, thus resulting in financial savings. »
The purpose of this organisation is to create specifications to define the mathematical
functions the most frequently used in numerical algorithms. These algorithms
can then be integrated in applications specific to a number of activity fields
such as industrial control, the development of embedded software and numerous
research fields. They can therefore be easily transferred between researchers
and engineers in industrial and academic fields.
« By using the industrial standards developed by the Numerical
Mathematics Consortium, students can create algorithms compatible with
functions shared by traditional tools and be assured that their work will
be correctly integrated in other mathematical environments, states Robert
H. Bishop, Professor and President of the Aerospatial Engineering and
Mechanics Department for the University of Texas. Furthermore, with a
proven standard for numerical mathematics, I can be sure that my students
will be trained on tools and approaches that they will encounter in the
industry. »
The Numerical Mathematics Consortium will create a mathematical community
within which the exchange of ideas and information will be facilitated by a
shared vocabulary.
The Numerical Mathematics Consortium is a non-profit organisation made up of
industrial companies and persons from the industrial sector and universities
with a view to defining an open standard for the semantics of mathematical functions
used in the development of numerical algorithms.
The main objective of the Consortium is to reduce the global costs of developing
algorithms and facilitating their use in a wide range of disciplines and in
various hardware and software environments.
Promoting the use of formal methods for software security, in particular within
the context of small secure portable objects and ubiquitous computing, is the
main goal of the INRIA EVEREST research project.
By offering Globalplatform, world leader for the development and deployment
of secure smart card solutions, a formal model for the GlobalPlatform
Card Specification v2.1.1 specification, the research project provides
an IT solution designed to reinforce the security assessment of industrial
products, accelerate the process for compliance testing and facilitate
the development of subsequent versions of GlobalPlatform products. Boasting
57 members that include Visa and Mastercard International, IBM, Hitachi,
Thales, STMicroelectronics, Sun, Gemplus, etc., the representatives of
this consortium stress the added value of this type of model that provides
« a rigorous and unambiguous specification » for those
wishing to implement the v2.1.1.
Formal verification techniques will thus contribute to a deeper understanding
of complex security architectures that are widely used in the industry.
Emerging from fundamental research, it will also increase confidence in
the solutions implemented by encouraging adoption of the most pertinent
standards.
© INRIA - 2005.