InterCor Hybrid-Roadmap v1.0 Final
InterCor Hybrid-Roadmap v1.0 Final
InterCor Hybrid-Roadmap v1.0 Final
communication
CONTROL SHEET
Version history
Name Date
Circulation
INEA 31/07/2019
Project Coordinator
Ronald Adams
Rijkswaterstaat
Legal Disclaimer
The information in this document is provided “as is”, and no guarantee or warranty is given that
the information is fit for any particular purpose. The content of this document reflects solely the
views of its authors.
The InterCor consortium members, jointly or individually, shall have no liability for damages of
any kind including, without limitation, direct, special, indirect, or consequential damages that
may result from the use of these materials.
Neither the European Commission nor the Innovation and Networks Executive Agency (INEA)
are liable for any use that may be made of the information contained therein.
TABLE OF CONTENTS
Control sheet ............................................................................................................. 2
Table of contents....................................................................................................... 4
List of Figures............................................................................................................ 5
List of Tables ............................................................................................................. 6
Acronyms and abbreviations ................................................................................... 7
1 Executive summary .......................................................................................... 10
2 Introduction ....................................................................................................... 11
2.1 Purpose of this document.......................................................................................11
2.2 InterCor Contractual References ............................................................................12
7 Technical challenges........................................................................................ 38
8 Summary and Conclusions.............................................................................. 39
9 References ........................................................................................................ 40
LIST OF FIGURES
Figure 2.1: High level diagram for hybrid communication from [8]. .......................................11
Figure 4.1: General system overview for the functional usage of IF2. ...................................18
Figure 4.2: Model A: deployment model based on national nodes that only serve national data.
The bold lines are implementations of IF2, where the broker and client boxes refer to the broker
and client implementations of IF2. The arrows indicate the direction of the data streams for the
current services. ...................................................................................................................19
Figure 4.3: Model B: deployment model based on national nodes exchange all data. The bold
lines are implementations of IF2, where the broker and client boxes refer to the broker and
client implementations of IF2. The black arrows indicate the direction of the data streams for
supporting vehicle from country 1 in country 2, and the orange arrows for vehicles from country
2 in country 1........................................................................................................................20
Figure 4.4: Model C: deployment model based on a central EU node. The bold lines are
implementations of IF2, where the broker and client boxes refer to the broker and client
implementations of IF2. The black arrows indicate the direction of the data streams for
supporting vehicle from country 1 in country 2, and the orange arrows for vehicles from country
2 in country 1........................................................................................................................21
Figure 5.1: High level overview of the organisation of the European C-ITS trust model. .......27
Figure 6.1: Functional diagram of the French hybrid architecture. ........................................32
Figure 6.2: Messages exchanged between the Nfr-ITS-S and other servers. .......................33
Figure 6.3: Definition of the technical interface for the harmonized French hybrid C-ITS
architecture. .........................................................................................................................34
Figure 6.4: Schematic overview of the systems used in the Talking Traffic project for signalized
intersection based services. .................................................................................................36
LIST OF TABLES
Table 4.1: Assessment of the three deployment models.......................................................21
Table 6.1: Summary of the communication stacks used on the interfaces of the French
harmonized hybrid architecture. ...........................................................................................34
1 Executive summary
InterCor has developed, implemented and tested the specifications for information exchange
via back-office systems enabling internationally interoperable hybrid communication. To
progress the usability of these specifications, several aspects need to be further defined and
tested. This document addresses two important aspects outlined hereafter.
The specifications of IF2 (i.e. the back-office interface) leave it open how these specifications
will be implemented. Three implementation models are discussed for the implementation of
IF2. These models mainly differ on how data is aggregated and distributed amongst Member
States: services providers can obtain the data from the country of origin, a single point of
contact per country can aggregate all data from other Member States and forward it to its local
service providers, or a single central aggregation point can be established where all data from
all Member States comes together, and is redistributed to all services providers. The models
are not mutually exclusive, and mixed forms can also be implemented. Various advantages
and disadvantages have been discussed. Within InterCor, no decision has been made on what
is the preferred implementation model.
The second important aspect covered in this document is how trust in the messages distributed
via IF2 can be established. For this aspect, two trust models are presented and discussed that
can be used to establish the required level of trust in cellular communication in the context of
hybrid communication. Descriptions of these models are provided, including some of the
identified strengths and weaknesses. Practical examples of two implementations have been
described as well. Existing security and communication standards need to be further optimized
in order to support an interoperable and harmonized solution over IF2 to deploy hybrid based
C-ITS services in the context of the InterCor project. Therefore, no final decision is taken on
the preferred trust model to use for IF2. What ultimately is required is a solution that is
optimised for hybrid communication, where stations can either implement only ITS-G5
communication, only cellular communication, or a combination of both.
2 Introduction
One of the objectives of InterCor is to provide C-ITS services on a broader scale by specifying,
using and advancing a hybrid communication approach to utilise a combination of cellular
(network-based) communication and direct (localized) communication. InterCor focuses on
realizing international interoperability of C-ITS services based on hybrid communication
solutions. In contrast, InterCor does not focus on the improved quality of service itself by
combining multiple networks, but extending the geographical availability of services for end-
users by connecting hybrid solutions in different countries.
Hybrid communication in InterCor is further reduced in scope by only considering ITS-G5 and
currently available cellular technologies: other communication technologies and future
extensions of these two technologies are not (explicitly) considered.
Based on this scope, InterCor has developed, implemented and tested the specifications for
information exchange via back-office systems [1, 8] enabling internationally interoperable
hybrid communication.
IF 2: Back-office interface
Service Service
Data Traffic Data Traffic
provider management provider management
provisioning provisioning
backend backend
Vehicle / Vehicle /
IF 3: Cellular interface
Personal Personal
system system
Figure 2.1: High level diagram for hybrid communication from [8].
To progress the usability of these specifications, several aspects need to be defined and
tested. This document addresses two important aspects. The specifications of IF2 leave open
how these specifications will be implemented. Several options can be considered. These are
elaborated and discussed in more detail is section 4. Another important aspect is how trust in
the content of messages distributed via IF2 can be established, which is the topic of section 5.
To make these aspects more concrete, section 6 provides two examples of concrete
implementations where these aspects have been addressed and specific solutions have been
chosen.
InterCor (Interoperable Corridors) links the C-ITS corridor initiatives of the Netherlands (among
which the C-ITS Corridor Netherlands-Germany-Austria), the French (among which the one
defined in SCOOP@F) and extends to the United Kingdom and Belgium C-ITS initiatives.
InterCor is an action co-financed by the European Union under the Grant Agreement number
INEA/CEF/TRAN/M2015/1143833. The Project duration is 36 months, effective from the 1st of
September 2016 until the 31st of August 2019. It is a contract with the Innovation and Networks
Executive Agency (INEA), under the powers delegated by the European Commission.
Any communication addressed to the Agency by post or e-mail shall be sent to the following
address:
For submission of requests for payment, reports (except ASRs) and financial statements:
[email protected]
Any communication addressed to the Agency by registered mail, courier service or hand-
delivery shall be sent to the following address:
https://webgate.ec.europa.eu/tentec/
All communication with the INEA or the European Commission shall be done via the Project
Coordinator, Mr. Ronald Adams.
The InterCor project has to be considered in the context of many ITS and C-ITS research and
development initiatives carried out across Europe over the last three decades. In this section,
the European C-ITS innovation process is sketched, and the position of InterCor in this
process. To identify its influence, it is important to recognize its contribution to the overall
European C-ITS objectives, relations with other initiatives and how it can be placed in C-ITS
innovation process.
According to the Oxford English dictionary, innovation means introduction of something new.
The National Innovation Initiative (NII) of USA defines innovation as the intersection of
invention and insight, leading to the creation of social and economic value. Dr A.P.J. Abdul
Kalam, the president of the Indian National Innovation Foundation pronounced “Innovation is
the celebration of creativity”.
From a process perspective all the definitions direct in a practical view “Innovation = Invention
+ Social-Economic Exploitation”. The innovation process is the translation of a new invention
into something socio-economically valuable or, in other words, it is to bring research into
practical socio-economic value.
Innovation is mostly a bottom-up process, but legislation may well take the initiative by ruling
societal or economic requirements (such as the reduction of CO2 emissions). There are
different types of innovations: Business, Architectural, Modular, Incremental, Radical and
Novel, all having different effect on the existing situation. Differences we need to consider in
the general development of C-ITS. The innovation process involves search and selection of
technologies, techniques, processes and business approaches. It requires exploration and
synthesis, cycles of divergence and convergence. Innovation needs at least the support from
several organisational levels, and in some cases deep involvement of these levels.
In the case of C-ITS, one can recognise three innovation levels as depicted in Figure 3.1.
As can be seen in the case of C-ITS, a large number of authorities and stakeholder’s groups,
represented by business organisations, are involved. This large number of groups and
organisations results in many organized and bilateral alignments, which are needed to come
to the realisation of an acceptable level of interoperability of an innovative solution for all the
stakeholders across all Member States.
To come to a common result, the agreement of the objectives and processes is an alignment
process in itself. Starting the innovation cycle, functionality and technology generally are
developed in bottom-up research processes. Investment in research starts slowly, and at the
moment benefits are more visible as investments increase. While based on new service
interests and introduction of new technologies, new innovation cycles are identified and
started. Each of these new cycles, however, need to go through the same steps to get to
maturity. The figure below identifies these steps for C-ITS.
Figure 3.2: The Innovation steps in C-ITS. Note that a complete overview of all
involved activities is not included.
In this innovation process, 4 stages can be identified. In the first stage, initial research is
performed on a wide range of technologies to determine an initial set of candidate solutions
for specific C-ITS services. In the second stage, the focus is on innovation, i.e. combining
technology with socio-economic exploration to come to realistic solutions that can be deployed
from both a technical perspective and a socio-economic perspective. In the third stage, initial
deployments are realized to gain field operational experience with these solutions. During
these initial deployments, lessons learnt on how to best deploy the services are used to fine-
tune the overall solution for deploying the services in a European wide context. After this 4-
step process, the services will become a commodity where all stakeholders have a common
understanding on their role in the service delivery, and how to fulfil that role.
A number of ITS related services (also) using standard cellular communication technologies
for their information sharing are commodity, i.e. the have reached the end of the above
sketched innovation cycle. An example is the route-planner service provided by navigation
systems providers. They adopt the route based on shared traffic information provided e.g. via
TMC or a dedicated connection to the back-office of the navigation system provider.
For the exchange of safety related information for C-ITS services, technology based IEEE
802.11p and ETSI ITS standards was chosen after the Initial research phase. Today we are at
the initial deployment stage, which is necessary to ensure interoperability and service
availability, also across country borders.
The introduction shows where the ITS community stands in the C-ITS innovation process. The
initiation of the C-Roads Platform (www.c-roads.eu) and implementing projects (such as
InterCor) proved to be a successful approach bringing European Member States interested in
C-ITS to real deployment. The development of C-ITS roadmaps was started by the Amsterdam
Group and then brought to a higher level by the European Commission. The EU C-ITS
Deployment Platform in its Phase-1 (2014-2016) and Phase-2 (2016-2017) led to the
formulation of the first EU vision and roadmaps [10]. The reports identified Day-1, Day-1.5 and
Day-2 services. To realize interoperability for these services, alignment between different
stakeholder groups, C-Roads Platform, C2C-CC (www.car-2-car.org) and others is necessary.
For instance, C-Roads and C2C-CC are developing their own specifications and aligning those
with each other. As these groups are more and more working together, they organize common
meetings to align at an early stage. It can be expected that the speed and quality of the
interoperable specifications will increase as a result of the cooperation within the C-Roads
Platform and between C-Roads and C2C-CC.
The projects under the umbrella of the C-Roads Platform have and will start and end at different
moments. To consolidate the results of these projects in a commonly agreed solution, the C-
Roads Platform needs to be a continuous activity, and thus continue also after all projects have
ended. To find the approach how to continue after all projects are ended is a task for the C-
Roads Platform to realize in cooperation with the European Commission. The aim is to realize
a single platform for the harmonisation and interoperability of C-ITS services at European level
for all stakeholders. To make the overall process faster and more efficient, the support and
stimulation of a continuing independent process where all stakeholders can realize
interoperability neutrally should be stimulated for the continuity of the deployment of current
and new C-ITS and CCAM services.
InterCor is initiated by 4 MS with funding from the EC, positioned as an initial deployment of
Day-1 services, i.e. stage 3 of the innovation process. In that context, it provides and aligns its
finding to the C-Roads Platform. C-Roads uses these findings to consolidate the specifications
to increase the functional and cross-border interoperability of the Day-1 services and use
cases. The work done in InterCor on harmonization of message definitions, ITS-G5 based
implementations and PKI aspects are part of this work. These activities resulted in detailed
profiles and other technical documents validated through TESTFESTs and operational pilots,
whose outcomes have been provided to the C-Roads Platform for wider harmonisation.
At the same time, InterCor also addresses technical challenges for cross-border
interoperability that are not yet in the third stage of the innovation process. This is particularly
true for hybrid communication, which did not conclude the first two stages of this process (i.e.
research and innovation). For hybrid communication, InterCor focusses specifically on the
innovation step by bringing results from previous (European and national) research activities,
but also considering market development.
As mentioned section 2.1, various approaches / solutions have been investigated in relation to
implementation models and security models for hybrid communication1. The next sections
present the various solutions considered and a preliminary assessment, but no position is
expressed in relation to their adoption: no definitive decisions on these topics are made within
InterCor. A hybrid communication roadmap is identified, leaving the next steps to C-Roads and
other activities.
1
For hybrid communication, InterCor only focusses on the international interoperability, but not
on a complete solution. Therefore, these aspects are only addressed in such perspective.
4.1 Introduction
The Figure below provides a simplified functional architectural view of the use of IF2 defined
within InterCor, as described in [1, 8], when a vehicle from country 1 is visiting country 2. In
this simplified view, IF1 (Wi-Fi air interface) and IF3 (cellular interface) are left out. This and all
other Figures in this document should be interpreted for multiple countries, multiple service
providers and multiple vehicles, but for simplicity, only 2 countries, 1 vehicle and 1 service
provider per country are shown. Data and service providers within each Member State may be
able to use IF2 for data exchange within a single country, but this is outside the scope of this
document.
Figure 4.1: General system overview for the functional usage of IF2.
The IF2 interface is identified as the interface between road authorities, data providers and
service providers, in any combination, to support data distribution via cellular IP-based
networks to satisfy the specific C-ITS services of interest to road operators and services
providers ensuring cross border service availability. The interface specifications are given in
[8]. At the current state the supported services within InterCor are RWW, IVS, PVD, and
GLOSA. Additional services are expected in near future, but it cannot be anticipated that the
current specifications support them. For the four InterCor services above, the data messages
originate from the road operator (or a third party on behalf of the road operator) or a vehicle,
and the information needs to be provided to a system in the vehicle to provide the service to
the driver of the vehicle or to the road operator. The receiving actor should be able to identify
who the data/service owner is and who provided (delivered) the data, if the (optional) header
on IF2 has been filled in.
The IF2 interface specifications themselves only specify the technical details of the interface
itself, but not the complete system in which the interface will be implemented. For the pilot
operation phase, a specific implementation model (possibly different for each InterCor Member
State) needs to be selected. The choice of a specific model for pilot operation will be influenced
also by what implementation model is foreseen in the long term, extending the services
deployment after the completion of InterCor. In the following section, three implementation
models are described. It is also possible to mix these models, however for clarity this document
focuses on the advantages and disadvantages of the separate models. An assessment of the
models will be given in section 4.3.
4.2.1 Implementation model based on national nodes for national data only
In this model A (Service Provider (SP) oriented), every country has a national node that
services only national data (i.e. C-ITS related information). Service providers from different
countries all connect directly to the broker in this national node. Every national node will have
to serve all service providers with vehicles in their country, and each service provider will need
to connect to brokers in all countries where it has customers. This situation is depicted in the
figure below, with the national node depicted as ‘broker’ for data provisioning and interfaces to
a ‘client’ in service provider back-end systems in each country, marked with black and orange
arrows.
In this model, there is no need to connect the different national nodes to each other.
Figure 4.2: Model A: deployment model based on national nodes that only serve
national data. The bold lines are implementations of IF2, where the broker and client
boxes refer to the broker and client implementations of IF2. The arrows indicate the
direction of the data streams for the current services.
In this model B (national node (NN) oriented), the NNs are interconnected with each other, and
the service providers will (only) need to connect to a single NN. Every NN will connect to every
other NN and processes all data from the other NNs to to serve it to the connected service
providers2. This situation is depicted in the figure below.
Figure 4.3: Model B: deployment model based on national nodes exchange all data.
The bold lines are implementations of IF2, where the broker and client boxes refer to
the broker and client implementations of IF2. The black arrows indicate the direction
of the data streams for supporting vehicle from country 1 in country 2, and the orange
arrows for vehicles from country 2 in country 1.
In this model C (EU central node (CN) oriented), the national nodes all forward their data to a
central EU broker, and all service providers connect to the central broker node3. The EU central
node can connect as a client to all brokers in all national nodes, and the service providers can
connect as client to the broker in the EU central node. This situation is depicted in the figure
below.
2
This implies that service providers only connect to a specific national node, which might not
be the case for internationally operating service providers. For simplicity, this is not taken into
account here.
3
Possibly they also connect to their national node, but that is out of scope of InterCor.
Figure 4.4: Model C: deployment model based on a central EU node. The bold lines are
implementations of IF2, where the broker and client boxes refer to the broker and
client implementations of IF2. The black arrows indicate the direction of the data
streams for supporting vehicle from country 1 in country 2, and the orange arrows for
vehicles from country 2 in country 1.
The three models described in the previous section, will be assessed based on the technical
complexity for the service provider and national node implementations, for their scalability, and
on their organisational complexity.
Depending on how strict the models will be implemented by different Member States, it is
possible to have deployments based on the different models and still have international
interoperability. Model A is the base model or starting point that needs to be supported by
4
See next section on a more detailed discussion of “trust”, which can also help to overcome
this legal issue.
individual Member State to provide data from road authorities on a central level per Member
State.
The models B and C can be deployed in time and are based on model A. It should be noted
that pan-European service providers are not tied to a single country and are free to connect to
multiple national nodes. The aggregation of data by national nodes (supported by local
government) can be offered by either each national node (model B) or by a central node (model
C). The functionality of the EU central node may also be provided by commercial providers
e.g. data aggregation providers to other service providers. For example, if a MS implements
model B in addition to model A for their existing service providers, i.e. provide data from their
own country and also from other selected Member States, via IF2 to their existing service
providers, this will be at the same time attractive to provide the data from that Member State
to new a service provider.
Multiple models based on model A will likely evolve and can be deployed and supported at the
same time. Consequently, it will become more complex to figure out what information can be
obtained where. Also, if information will be provided at different places, service providers
subscribing to this information need to ensure that they are able to identify duplicates and
ensure they can handle that correctly.
Therefore, it might be preferable to have a common approach for selecting one of the
deployment models B or C, at least for the long term. Support for multiple models could be a
good way to start off, as long as it is unclear what that common approach should be and
migrate in a later stage. Also, model C can be treated as a development of model A or B to
simplify the data distribution on a larger (European) scale. The move from model A to B or C
will depend on the benefits for both national nodes to distribute data and for services providers
to collect data (and vice versa for probe vehicle data). The EC could stimulate the use of pan-
European ITS safety services by giving support to an EU node.
When scaling up to a larger deployment, it can be considered to add a control layer to the
currently defined operational layer. Such a control layer would facilitate dynamic requests of
technical details about where to get access to what data. This could be implemented as a kind
of market place, as investigated in e.g. the Mobinet project. Another option would be to
standardize a control layer where every application that wants to connect via IF2 first
addresses a central server with the details of the information that is required/provided
(message type, geographical region of interest, etc.), and based on that receives the details
(IP address, credentials, etc.) of an IF2 implementation to connect to. This approach has been
implemented already in the Talking Traffic project. When a market place or control layer would
be added, it is not necessary anymore to make a single, EU-wide decision on which of the
models, or combination of models, will be supported.
For the testing and piloting phase in the frame of InterCor, the participating Member States
(France, Belgium/Flanders, Netherlands, and the United Kingdom) have chosen to support
multiple implementation models to realize international interoperability. This does not reflect
the preferred long-term solution. At this moment, InterCor does not provide a preferred
implementation model in the long-term perspective. This issue needs to be addressed in a
larger context, as consensus is required on a European level. This issue could therefore be
further addressed in the C-Roads Platform and/or other European C-ITS deployment projects.
An important aspect for message exchange over IF2 is how to ensure trust. Ultimately, the
system receiving messages needs to be able to verify the authenticity and integrity of the
received messages. Only if the receiving system trusts the content of a C-ITS message, it can
use this message in the services it provides.
To establish trust within a particular domain, agreements, processes and procedures need to
be established and translated into policies and standards. These put legal and technical
requirements to all participants within the trust domain. Specifications are essential as they
clearly specify the requirements on messages that are transmitted in the context of the trust
domain. These requirements are covered in standards, and complemented by profiles, which
are currently being investigated by a.o. InterCor, the C2C CC and the C-Roads Platform. Also,
a compliance assessment procedure shall be in put place to ensure that the standards and
profiles are properly implemented by transmitting systems, but this is out of the scope of
InterCor and will not be treated here.
For messages exchanged via IF1 (i.e. via ITS-G5), a European C-ITS Trust model has been
developed [11], defining the trusted C-ITS communications system. This model is based on
Public Key Infrastructure [12] within the scope of the overall EU C-ITS Security Credential
Management System (EU CCMS).
A complicating factor is that the trust needs to be ensured for specific end-to-end services,
using a specific communication channel, or even for a complete architecture, whereas InterCor
only defines the interoperability interfaces, and not the overall architecture or communication
chain. Therefore, only an overview of possible solution directions is provided and assessed
here in the context of IF2.
In the next sub-section two trust models are described and discussed. Although they are
discussed separately, in practice they can also be used in combined fashion.
Early 2014 the European Commission decided to take a more prominent role in the deployment
of connected driving by setting up a C-ITS Deployment Platform [10]. The EU C-ITS Platform
Among other aspects, the security of V2X communications has been covered by the EU C-ITS
Platform. A security architecture with support of a Public Key Infrastructure (PKI) using
commonly changing pseudonym certificates has been eventually developed [11, 12].
The common C-ITS Certificate Policy [13] defines the European C-ITS trust model based on
Public Key Infrastructure. It defines legal and technical requirements for the management of
public key certificates for C-ITS applications by issuing entities and their usage by end-entities
in Europe. The PKI is composed at its highest level by a set of Root Certificate Authorities
(RCAs) “enabled” by a Trust List Manager (TLM), i.e. whose certificates are listed in a
European Certificate Trust List (ECTL) defined and published by the central entity TLM.
Figure 5.1 provides a pictorial description of the C-ITS trust model architecture. The Policy
Authority appoints the TLM and therefore provides trust in the operation of the TLM to all PKI
participants. The Policy Authority approves the Root CA operation and confirms that the TLM
can trust the Root CA(s). The TLM issues the ECTL that provides trust in the approved Root
CAs to all PKI participants. The Root CA issues certificates to the Enrolment Authority (EA)
and Authorisation Authority (AA) and hence provides trust to their operation. The EA issues
Enrolment Certificates to the sending and relaying ITS station (as End-Entity), providing trust
in its operation. The AA issues Authorization Tickets to the ITS stations based on the trust from
the EA.
Figure 5.1: High level overview of the organisation of the European C-ITS trust model.
The C-ITS trust model is based on a multiple Root CA architecture and is composed of the
following main components:
• The Policy Authority: The Policy Authority is a role composed by the representatives
of public and private stakeholders (e.g. Member States, Vehicle Manufacturers, etc.)
participating to the C-ITS trust model, where a consensus based voting scheme
applies. The Policy Authority is the main decision body that guarantees the trust and
interoperability between the participating members states PKIs.
• The Trust List Manager (TLM): The TLM is a unique entity appointed by the Policy
Authority. The Trust List Manager is responsible for operation of the ECTL according
to the common valid CP and regular activity reporting to the policy authority for the
overall secure operation of C-ITS trust model.
• The C-ITS Point of Contact (CPOC): It is the responsibility of the Policy Authority to
authorize the CPOC to fulfil some main tasks to secure communication exchange
between all entities of the C-ITS trust model including the transmission of the Root CA
certificates to the TLM as well as the publication of the ECTL.
• The Accredited Auditor: The Accredited Auditor is responsible for performing or
organizing audits of root CAs, notifying to the Policy Authority on the successful or
unsuccessful execution of an audit either initial or periodic and assessing compliance
of CPSs to the CP.
• Members States PKIs elements: Each Member State is responsible of managing its
own PKI composed of Root CAs, EAs and AAs. The certificates of the Root CAs are
transmitted periodically to the CPOC through a secure protocol (which is under
construction by the EC).
To apply this model for IF2 as it has been defined for ITS-G5 Day-1 services, the following
requirements need to be respected:
• A sending ITS station acquires rights to broadcast an authorized set of ITS messages
from Enrolment Authority, negotiates rights to invoke ITS services from Authorization
Authority and sends single-hop and/or relayed signed broadcast messages.
• A relaying ITS station (Forwarding ITS station) receives broadcast messages from the
sending ITS-S and forwards them to the receiving ITS station if required.
• A receiving ITS station receives signed broadcast messages from the sending or
relaying ITS station.
• Following the certificates management procedures as they have been defined in the
Certificate Policy (CP) in terms of time of validity of certificates, signature algorithms
and cryptographic requirements, verification of signature, generation of keys for
signature and encryption, etc.
Advantages of this model are that it can be integrated relatively easy with ITS-G5 based
communication, that it does not put strong trust requirement on any relay station, and that it is
already accepted by many stakeholders in the C-ITS domain. However, several difficulties
have been encountered when applying this model to cellular technology:
• The current C-ITS security standards and EC CP have been mainly focused so far on
ITS-G5 access technology; hence the signature has been set at the network layer,
necessary for this type of technology. However, for IP based links based on another
transmission paradigm, carrying message signature at the Geonet layer may appear
as not relevant and restrictive.
• End-to-end data message signature is required for trust and message integrity and
message authorisation verification; this must be kept, but maybe without the complete
structure of Geonet Secured Message for C-ITS based cellular communications. ETSI
is working on a ITS standard to allow message signatures on the facility layer or
application layer instead of the Geonet layer. This could be useful for long-range,
cellular-only communications.
• In order to provide service interoperability between Member States, harmonisation
should be done to integrate into one solution different signing approaches (occurring at
facility layer, at network layer or at application layer) taking into account the constraints
of communicating end user vehicles performance (hardware security modules cannot
sign many messages per second) as well as C-ITS services performance constraints.
• Another difficulty on the hybrid link is that the information can be more widespread than
on ITS-G5 and the reciprocity of the exchange of information might be requested by
different stakeholders. However, the European C-ITS trust model allows a “club of
senders” to send messages but doesn’t prevent others outside the club to receive and
use the messages.
A common model to realize trust in the context of internet services is via a so-called chain of
trust. In this model, information is exchanged between peers that have an established relation
with each other. For every relation, it is agreed what information is exchanged, and what
standards and profiles are adhered to. Part of the agreement will also include how peers’ proof
that they adhere to these standards and profiles. In the operational phase, at a minimum, it is
required that it can be established that the communication is taking place with the correct peer,
i.e. the peer is identified in the operational phase.
In the complete implementation of a service, information can flow via systems of multiple
entities, where every peer to peer communication is governed by an established relation with
a related level of trust. In this way, a chain of trust is formed between the originating source of
information and the final destination. As a baseline, the level of trust will generally only
decrease when it flows via the chain. However, it is possible that a node in the chain can
increase the level of trust by e.g. combining information from multiple sources, or by other
means to validate the information in the messages.
The peer-to-peer trust relations can be established in many ways and at different layers in the
OSI communication stack. For the end-to-end trust, it is not required that the same mechanism
is used at every peer-to-peer connection in the full chain. At the lowest layer, physically
secured network connections can be used, typically for connections within a single entity. At
higher layers, Virtual Private Networks (VPN), IP Security (IPSec), or Transport Layer Security
(TLS) can be applied. At higher levels in the OSI stack, authentication mechanisms based on
simple username/password authentication, OAUTH, or SASL can be used. In the end, based
on the specific technology and implementation details, a certain level of trust can be realized
between peers, and in that way end-to-end.
The advantages of this model is that technical solutions can be optimized based on specific
needs of specific stakeholders, and it is not required to agree on an overall solution with all
stakeholders before services can be implemented. This model is well accepted in the mobile
internet domain in general, but not yet applied to mobile C-ITS services specifically. As the
end-to-end trust is build-up of the trust on every peer-to-peer connection in the chain, every
intermediate (relay) node needs to be trusted as well: relaying via untrusted nodes is not
possible5. Furthermore, integration with ITS-G5 trust model to come to a truly hybrid model
might prove difficult.
For the chain-of-trust model, it is unclear how to convert specific technology solutions to trust
levels. This is required to determine the overall end-to-end trust level. Technical specifications
of the European C-ITS trust domain are to a large extent finished and implementation is
underway; the chain of trust has a lot of freedom for implementation. It could be useful to
investigate to what extent the architecture, trust model, and/or technical specifications of the
5
In this context, only nodes that operate on the level where the trust is technically implemented
are considered; so, if the trust is realized on the application level, intermediate nodes on the
network level (e.g. routers) are not considered to be part of the chain of trust.
European C-ITS trust domain as developed for ITS-G5 communication could be used also for
the chain of trust. For example, it could be investigated whether the Certificate Policy of this
trust domain can be used when providing certificates that are used in VPN or TLS
implementations. This would most likely require modifications of the Certificate Policy, but
would allow operating in the same trust domain, independently from whether individual
messages are signed based on TS 103 097 or the connection between peers is secured via
VPN or TLS.
For the French C-ITS projects (SCOOP@F, C-Roads France, InterCor), a complete and
harmonised C-ITS hybrid architecture has been developed. It makes sure that the same
message going through different channels of communication can be understood. The
functional links are shown in the figure below.
The Nfr-ITS-S is the French National ITS station. As any C-ITS station, it is enrolled in the
French PKI.
• Router for any C-ITS message, with notably geographical filtering functionality; it also
implements the IF2 to communicate with any back-end server, including foreign C-ITS
stations;
• Receiver of Datex II messages and building C-ITS messages from them (functions for
road operators), like any roadside ITS station in France; it also works the other way
around; furthermore, it adds some specific road operator functions, such as the
aggregation of CAM messages.
The detail of the messages exchanged between the Nfr-ITS-S and other servers is shown in
the figure here below.
Figure 6.2: Messages exchanged between the Nfr-ITS-S and other servers.
• on the link itself: TLS 1.2 is used with the exchange of x509 certificates;
• for each message: messages are exactly the same ones signed on the ITS-G5 link
(including the Geonet header) within the C-ITS trust domain – the Nfr-ITS-S verifies
each message that arrives before forwarding it (or publishing it) on every link;
• on the non C-ITS links: specific bilateral solutions between the French Ministry and
road operators have been defined.
Figure 6.3: Definition of the technical interface for the harmonized French hybrid C-ITS
architecture.
The choice was made to secure C-ITS messages for the interfaces numbered 1, 2, 4, 6 and 7
by using signature at Geonet layer and to use DSMIP tunnels for IPv6.
This choice was made because for now the standards allow only such implementation. It was
also difficult for the vehicles to sign twice and some vehicles were directly connected to the
Nfr-ITS-S.
Table 6.1: Summary of the communication stacks used on the interfaces of the French
harmonized hybrid architecture.
1 ASN1 u-per / BTP / Geonet / TCP DENM IVI SPATEM CAM , DENM
/ IPv6 / cellular MAPEM POI
security at Geonet level
implementation of DSMIP tunnel
2 ASN1 u-per / BTP / Geonet / TCP DENM IVI SPATEM CAM and DENM
/ IPv6 / ITS-G5 MAPEM POI
security at Geonet level
3 Out of scope
6 ASN1 u-per / BTP / Geonet / TCP DENM IVI SPATEM CAM and DENM
/ IPv6 or IPv4 / ITS-G5 MAPEM POI
security at Geonet level
7 ASN1 u-per / BTP / Geonet / TCP DENM IVI SPATEM CAM et DENM
/ IPv6 or IPv4 / cellular MAPEM POI
security at Geonet level
implementation of DSMIP tunnel if
using IPv6 link
8 IF2 from InterCor DENM IVI POI DENM POI ETA
ASN1 u-per / AMQP / TCP / IPv4 information
security at Geonet level
9 IF2 from InterCor DENM IVI SPATEM DENM
ASN1 u-per / AMQP / TCP / IPv4 MAPEM POI
security at Geonet level
10 IF2 from InterCor SPATEM, MAPEM
ASN1 u-per / AMQP / TCP / IPv4
security at Geonet level
The Dutch project Talking Traffic has developed specifications and implementations C-ITS
services based on cellular communication. One set of services are related to signalized
intersections, see figure below.
IF2
Data Service
RIS TLEX
provider provider
Vehicle/personal
system
Figure 6.4: Schematic overview of the systems used in the Talking Traffic project for
signalized intersection based services.
A new architecture and interfaces have been developed for the so-called Intelligent Traffic
Light Controller (iTLC, or iVRI in Dutch). Without going into details of that architecture, a
relevant aspect is that the Roadside ITS Station (RIS) is functionally separated from the rest
of the traffic light controller. The interface between the RIS and the rest of the TLC is based on
IP communication, and is secured with TLS. The certificates used for this interface are provided
from a dedicated “TLC” trust domain. Access to this trust domain is granted after an established
compliance assessment process has been completed.
All RIS communicate with a central system called TLEX (Traffic Light Exchange), which is the
central node. The interface between the RIS and TLEX contains both a management protocol,
based on HTTPS/web services technology, and a streaming protocol for the exchange of
actual ASN.1 encoded, standardized messages. The link is secured based on TLS with generic
(internet) certificates and temporary authorization tokens. Access to TLEX is granted after
successful completion of a compliance and interoperability test program.
Multiple data providers can obtain the data from TLEX, based on a similar interface and trust
model as is in place between the RIS and TLEX, which includes: separate management and
data flow protocols, security implemented based on generic certificates and authorization
tokens. Also for this interface, access is granted based on the successful completion of a
compliance and interoperability test program6.
6
As an extension to the Talking Traffic specifications and implementation, an implementation
of IF2 has been added to provide the MAP and SPAT data for the realization of the InterCor
GLOSA service.
Data providers can enrich the information with data from other sources, and can provide
additional services to service providers, e.g. managing what information is relevant for what
end-user. The (business) relation between data providers and service providers (that have the
direct relation with end-users) is up to the data and service provider. Also the technical
interface and method to establish the trust relation is not prescribed. This is also true for the
message exchange between the service provider back-end and end-user device7 (vehicle or
personal system).
Via the same data path and trust relations, also information from the vehicles/personal systems
is communicated to the RIS connected to the traffic light controllers. For the two interfaces to
TLEX, CAM and SRM messages are exchanged towards the traffic light controllers.
In summary, the trust within this eco-system is based on agreed specifications, interoperability
and compliance test specifications, and various ways of identification and authentication on
link level for the various interfaces. The current specifications do not include signing of
individual messages. It is currently under investigation whether there is an added value in
(also) supporting the exchange of signed messages to increase the level of trust. This could
be used, e.g. for the SRM messages containing priority request. If and how this would be
realized is still under investigation as part of a possible future upgrade of the specifications.
7
Strictly speaking, it is not even required that the service provider has a back office that is part
of the real-time data flow; it is possible that the data provider delivers the data directly to the
vehicle or personal system, without (technical) intervention of a back office system of the
service provider.
7 Technical challenges
Several technical challenges still remain to realize end-to-end trust. For the application of the
European C-ITS trust domain, complete technical specifications do not exist yet. The current
specifications assume that messages are signed on the Geonet layer. The current and
previous versions ETSI TS 103 097 specify how a message included in a Geonet layer should
be signed, and how the required information to interpret the signatures can be exchanged. The
Geonet protocol is designed to be used on ad-hoc broadcasting networks. Although it is
technically possible to use other OSI layer 2 protocols than ITS-G5, the protocol has been
developed based on (implicit) assumptions of ITS-G5, like a strong link between message
reception and geographical distance, open broadcast (i.e. without encryption on lower layers),
and limited reliability. Furthermore, the purpose of the protocol is to be able to handle the
limitations of such ad-hoc broadcast networks. Therefore, it is unclear what the added value is
for the inclusion of a Geonet layer on the cellular link of a hybrid communication solution, and
what modifications to the Geonet standard are required to make it fully operational on cellular
networks. So, although inclusion of the Geonet layer has the advantage that specifications
exist on how to sign messages, the overall solution is not well defined, and it is not evident
whether it would be technically the most appropriate solution. This solution has been
implemented and tested in France.
The messages could also be signed at the facilities layer. The signing principles from TS 103
097 can also be applied to a payload that only contains the UPER encoded message, but the
details need to be worked out and tested in pilots. Furthermore, as this solution is not yet
standardized, it is not generally supported by existing systems implementing the European C-
ITS trust domain. Signing messages at different layers will require every message to be signed
twice in hybrid C-ITS stations, which might impact the performance.
A more practical technical choice is the version of the security standard to use. InterCor has
decided to use TS 103 097 version 1.2.1 for the IF1 interface (ITS-G5), and the use of the
same standard on IF2 would be a logical choice. Other European initiatives have already
decided to use the newer version of the standard, version 1.3.1, which would be more future
proof. The choice of the version is irrespective of whether the signing will be done on the facility
or Geonet layer. Although this is “just” a technical choice related to the moment in time (also
all InterCor Member States have expressed the intention to migrate to the new standards), the
issue of (backward) compatibility will remain relevant also in the coming future. As the context
of hybrid communication is more complex than for ITS-G5 stand alone, the issue of future
upgrades will also become more complex. A clear approach supporting future upgrades needs
to be developed, possibly before deployment.
Three models have been discussed for the implementation of IF2. These mainly differ on how
data is aggregated and distributed amongst the Member States back offices: services
providers can obtain the data from the country of origin (national node), a single point of contact
per country can aggregate all data from other Member States and forward it to its local service
providers, or a single central aggregation point can be established at European level, where
all data from all Member States comes together and is redistributed to all services providers.
These models are not mutually exclusive and mixed forms can also be implemented. Various
advantages and disadvantages have been presented. Within InterCor, no decision has been
taken on what model is the preferred implementation.
Two trust models have been presented and discussed that can be used to establish the
required level of trust in cellular communication in the context of hybrid communication.
Descriptions of these models are provided, including some of the identified strengths and
weaknesses. Practical examples of two implementations have been described as well. Existing
security and communication standards need to be further optimized in order to support an
interoperable and harmonized solution over IF2 to deploy C-ITS services based on hybrid
communication at European level. What ultimately is required is a solution that is optimised for
hybrid communication, where C-ITS stations can either implement only ITS-G5
communication, only cellular communication, or a combination of both.
9 References
[3] InterCor, “Milestone 6 Common set of upgraded specifications for Services v.2.0”, 2019.
[4] Project Talking Traffic, “RFP Talking Traffic 1.1_Bijlage 9 Latency Tabel_Beter
Benutten_2016.07.01.pdf”, 2016.
[6] ISO, “EN ISO 17423:2017 Intelligent transport systems -- Cooperative systems --
Application requirements and objectives”, 2017.
[7] InterCor, “Milestone 3 - Common set of upgraded specifications for ITS-G5 v1.1”, 2017.
[9] Igor Passchier, “InterCor Deployment models for IF2”, version 1.0, 5-4-2018.
[10] C-ITS Platform Phase I “Final report”, Jan. 2016; C-ITS Platform Phase II “Final report”,
Sep. 2017, both available at https://ec.europa.eu/transport/themes/its/c-its/.
[11] ETSI “TS 102 940 v1.3.1 Intelligent Transport Systems (ITS); Security; ITS
communications security architecture and security management”, 25-4-2018
[12] ETSI “TS 102 940 v1.3.1 Intelligent Transport Systems (ITS); Security; Trust and
Privacy Management”, Feb. 2019
[13] C-ITS Platform Phase II, “Certificate Policy for Deployment and Operation of European
Cooperative Intelligent Transport Systems (C-ITS)”, Release 1, June 2017,
https://ec.europa.eu/transport/sites/transport/files/c-its_certificate_policy_release_1.pdf