SALT Frameworks to Tackle Surveillance and Privacy Concerns
Antonio Kung, Christophe Jouvray and Fanny Coudert
Trialog, Paris, France
Keywords: PET (Privacy Enhancing Technology), AET (Accountability Enhancing Technology), SALT (Socio-
Contextual, ethicAL, Legal, Technical) Framework, Privacy-by-Design, SFMT (SALT Framework
Management Tool).
Abstract: This paper elaborates on the need to take into account the different views of the stakeholders involved in the
development of surveillance systems and civil society, during the design process. It first provides an
overview on privacy-by-design approaches. It then identifies three principles essential to integrate privacy
concerns into the design of surveillance systems. It consequently proposes a design process based on social-
contextual, ethical, legal and technical frameworks (SALT) and the challenges for its creation and use. It
finally provides a specification of a resulting SALT framework management tool based on modelling
techniques.
1 INTRODUCTION
Surveillance and Privacy are critical societal issues
today. In the wake of the Edward Snowden’s
revelations (http://en.wikipedia.org/wiki/Edward_
Snowden), several law proposals flourished in the
US Congress to increase the transparency and
accountability of the use by the government of its
surveillance powers (Electronic Frontier Foundation,
2013). Similarly in Europe, on 8 April 2014, the
European Court of Justice annulled the Data
Retention Directive that was imposing to Internet
Service Providers an obligation to retain all traffic
and location data generated by their communication
networks for purposes of investigation, detection and
prosecution of serious crime. One of the arguments
advanced by the Court is that such blanket data
retention obligation is likely to generate in the minds
of the persons concerned the feeling that their
private lives are subject of constant surveillance
(CJEU, 2014). These examples show that
surveillance and privacy are issues that, in a
democratic society, go hand in hand and must be
solved jointly. One way to do so is through Privacy-
by-Design, a concept that is turned into a legal
obligation for the development of new information
systems in the Data Protection Package proposed by
the European Commission in January 2012 (Reform
of the Data Protection Package, http://ec.europa.eu/
justice/data-protection/review/index_en.htm).
This paper elaborates on the need to develop an
adequate framework for the implementation of
Privacy-by-Design in the development of
surveillance technologies and systems.
2 TODAY’S PRACTICE OF
PRIVACY-BY-DESIGN
One of the big hopes in solving privacy issues
stemming from the development of new
technologies is the concept of Privacy-by-Design, a
term coined by Ann Cavoukian (Privacy-by-Design,
http://www.ipc.on.ca/english/Privacy/Introduction-to
-PbD/). Applied to the design of Information and
Communication Technologies (ICT) based
applications, Privacy-by-Design (PbD) focuses on
requirements and measures that take into account the
respect of the individuals’ privacy. The full
integration of PbD in today applications, systems
and development process is being worked out both
on the management front and the engineering front.
At the management level PbD is often associated
with Privacy Impact Assessments (PIAs) (Wright
and Hert, 20102). A PIA is defined as a systematic
process for evaluating the potential effects on
privacy of a project, initiative or proposed system or
scheme and finding ways to mitigate or avoid any
adverse effects (PIAF, http://www.piafproject.eu).
665
Kung A., Jouvray C. and Coudert F..
SALT Frameworks to Tackle Surveillance and Privacy Concerns.
DOI: 10.5220/0005444306650673
In Proceedings of the 3rd International Conference on Model-Driven Engineering and Software Development (SPIE-2015), pages 665-673
ISBN: 978-989-758-083-3
Copyright
c
2015 SCITEPRESS (Science and Technology Publications, Lda.)
At the engineering level, principles discussed in
(Spiekermann and Cranor, 2009); (Gürses et al.,
2011); (Kung et al., 2011) are used as an input to the
PRIPARE FP7 project to define a comprehensive
methodology (http://pripare.eu/).
Standardisation activities are also on-going.
Within ISO (ISO/IEC JTC1/SC27:
http://www.jtc1sc27.din.de), standards are being
prepared on PIAs
1
, and on a code of practice for
personally identifiable information protection
2
.
Within OASIS (https://www.oasis-open.org/), two
standards are available, PMRM (Privacy
Management Reference Model and Methodology)
(OASIS, https://www.oasis-open.org/committees/
pmrm/charter.php), which explains how privacy
principles are mapped onto operational requirements
(e.g. agreement, security, access) and PdB-SE
(Privacy by Design Documentation for Software
Engineers) (OASIS, https://www.oasis-
open.org/committees/pbd-se/charter.php), the goal
of which is to provide privacy governance and
documentation standards to software engineers.
Finally, the European Commission is in the process
of issuing a mandate for the establishment of
European standards on PbD (http://ec.europa.eu/
DocsRoom/documents/5290).
The question is, how can we support PbD in the
design of a surveillance system? Is it sufficient to
rely on generic PbD approaches? Do we need
specific features? Do we need standardisation? The
next section lists three principles that are needed in a
design process integrating PbD.
3 PRINCIPLES FOR PBD FOR
SURVEILLANCE
3.1 Multi-stakeholder Empowerment
The deployment of a typical surveillance system
usually involves a number of stakeholders. An
authority stakeholder (often public) would decide on
the deployment of a surveillance system. A
surveillance system owner would be responsible to
deploy and operate the system. A surveillance
system designer would be mandated to design the
surveillance system.
As more sophisticated surveillance capability
will be available in the future, and with the rising
concerns about privacy, additional stakeholders are
likely to gain influence in the deployment of
1
ISO 29134
2
ISO 29151
surveillance systems. Compliance of the system with
the legal framework will not only be subject to the
scrutiny of data protection authorities but also
internally to the one of data protection officers and
externally to third parties certifiers/auditors. Beyond
legal compliance, public opinion will also have to be
taken into account as it conditions the public
acceptance of the surveillance. Public opinion is
usually voiced either indirectly by privacy advocates
(e.g. privacy associations, or privacy activists) or
directly expressed on the internet (e.g. on social
networks) or through specific channels opened for
engaging a dialogue with civil society (e.g. meetings
with the affected community of citizens). Ethical
aspects of surveillance should be given more weight
to solve the most difficult issues not resolved by the
legal framework.
Furthermore, the rapid evolution of technology
and its profound societal impact implies that
technology makers and social sciences analysts will
also have a strong influence. At the technology level
more sophisticated privacy enhancing technology
(often called PETs) will be available, but likewise
more sophisticated surveillance technology will also
be available.
It follows that various forces will decide on the
fate and shape of a resulting surveillance system.
Such forces can be structured into three viewpoints:
(1) the socio-contextual ethical viewpoint, (2) the
legal viewpoint and (3) the technology viewpoint.
The PARIS project (PARIS, http://www.paris-
project.org/) has coined the term SALT or Socio-
contextual ethicAl Legal Technical to qualify the
framework.
The resulting principle is multi-stakeholders’
empowerment, i.e. each viewpoint should be entirely
taken into account in the design process.
3.2 Concerted Impact Assessment
As mentioned above, a Privacy Impact Assessment
or PIA (Wright and Hert, 2012); (PIAF,
http://www.piafproject.eu/) is a process for
evaluating the potential effects on privacy of a
project, initiative or proposed system or scheme and
finding ways to mitigate or avoid any adverse
effects. A risk analysis is carried out and if resulting
privacy risks are beyond some level, measures are
identified and implemented to eliminate or minimise
those risks. For instance (CNIL, http://www.cnil.fr/
fileadmin/documents/en/CNIL-ManagingPrivacyRis
ks-Methodology.pdf) is a privacy risk analysis that
is promoted by the French data protection authority.
Likewise, it is through an assessment of the
MODELSWARD2015-3rdInternationalConferenceonModel-DrivenEngineeringandSoftwareDevelopment
666
security risks that may exist in a given area that
recommendations for surveillance measures may be
taken. We could say that a security impact
assessment is carried out. It identifies the impact that
no measures would have on the security of the given
area (security of the elements that are in the area,
e.g. citizens, assets) and consequently sets out the
appropriate surveillance recommendations to
prevent crimes and terrorism.
Both privacy and security risks analysis should
be made in a concerted way. It is well known that
while surveillance measures have a negative impact
on privacy, privacy measures may decrease the
efficiency of surveillance measures. The only
solution so far is to carry out a proportionality
assessment in which the need and efficiency of the
claimed surveillance measures (potential benefit) is
weighted against their impact on privacy (negative
impact). In other words, the surveillance and the
privacy impact assessments are carried out jointly.
The resulting principle is concerted impact
assessment.
3.3 Trust through Accountability
Measures
The impact on citizens’ privacy of surveillance
measures that look for collecting comprehensive
information about potential criminals or terrorists is
not being discussed. Yet, citizens are willing to give
up part of their privacy to ensure a reasonable level
of security. A sufficient level of trust is therefore
needed between citizens and stakeholders in charge
of performing surveillance duties, i.e. in the fact that
the latter act within the powers granted to them. One
way to create trust is to establish sufficient
transparency and accountability. An overview of the
actions made by today civil society organisations to
monitor surveillance systems is provided by
(Surveillance, http://irissproject.eu).
Beyond regulation, transparency and
accountability can be made easier if accountability-
by-design is applied, i.e. accountability is a
requirement from the start and accountability
enhancing technology (AET) are used. Examples of
such technology are secure access logs, i.e.
technology that ensures that all access to data are
securely logged in such a way that it can be used for
accountability, e.g. for later judicial actions. Four
design attributes are defined in (Kung, 2014) that are
important in a privacy preserving system:
minimization of data, enforcement of data protection
policies, accountability and flexibility (to improve
the system). (Kung, 2014) also lists examples of
technology used for each attributes. Three types of
measures are associated with accountability: log data
access, log modifications (policies, level of
protection), and protect log data. With this type of
technology, data transmitted for surveillance
purpose to external parties but also internally would
be logged.
The resulting principle is trust through
accountability measures, i.e. all surveillance actions
are associated with measures that would later be
used either to prove that no privacy infringing
actions have been carried out, or to prove that a
privacy breach has happened.
4 A DESIGN PROCESS BASED
ON SALT FRAMEWORKS
Figure 1 describes the resulting design process based
on the above principles. It consists of three
components: a common reference or framework that
will be used by the stakeholders involved in the
process; the design process itself; the designed
system.
The common reference acts as a dedicated
repository of knowledge dedicated to a given
environment. This could be a geo-political
subdivision (e.g. city, a region, a country). It
combines information and knowledge concerning
several references, the Socio-contextual and ethicAl
reference, the Legal reference, and the Technical
reference. This is the reason why it is designated as a
SALT framework. The socio-contextual and ethical
reference contains parameters that are specific to a
region (e.g. France), and types of interactions (e.g. at
home, at work). The legal reference would relate to a
country’s legislation (e.g. Spain) or a given
technology (e.g. video surveillance). The technical
reference includes the wealth of available
technologies and practices. This would include
privacy enhancing technology (PET) and associated
practices. In the case of surveillance systems it
would further involve surveillance technology,
accountability enhancing technology (AET) and
associated practices.
The design process uses the common reference
to drive the design of a surveillance system design
that applies privacy-by-design principles as well as
accountability-by-design principles. Using the
common reference provides the assurance that all the
needed obligations, advised practices and
technology solutions are considered and integrated.
This is why it is called SALT design process.
SALTFrameworkstoTackleSurveillanceandPrivacyConcerns
667
Figure 1: SALT Design Process.
Once the design process has been applied, a
system is designed with privacy preservation and
accountability features. This is why it is called
SALT enabled or SALTED system. More
information can be found in (Deliverable D2.2,
http://www.paris-project.org/index.php/delivera-
bles).
5 SALT FRAMEWORK
CHALLENGES
There are several challenges in the use of SALT
framework that must be addressed: the capture of
knowledge; the need for different references; the
usage of a SALT framework; and the management
of a SALT framework. We briefly explain these
challenges and how we address them.
5.1 SALT Framework Knowledge
Capture Challenge
While the concept of a SALT framework is
powerful, it is only valid if the resulting repository is
comprehensive (i.e. all needed information is
stored), accurate (i.e. it reflects the knowledge of the
expert stakeholders) and flexible (i.e. if must cope
with specific information and allow for evolution).
Further, a SALT framework must be structured in
order to cope with the multidisciplinary needs.
Different viewpoints will be needed. For instance a
stakeholder with a legal interest will need very
precise access to legal references and a high level
access to technical references. Likewise, engineers
will need very precise technical references and high
level access to legal references. The challenge is to
structure information to provide different viewpoints
while remaining consistent (e.g. high level views are
consistent with detailed views). Another challenge is
to cope with terminology issues. Lawyers would be
at a loss if they were exposed to engineering
phraseology, and the same would apply to engineers
reading legal text.
5.2 Multiple SALT Frameworks
Challenge
There can be many SALT frameworks. A
surveillance system designed in France would have
to follow the obligations set out by the French legal
framework and the recommendations of public
bodies such as the CNIL, the French Data Protection
Authority (http://www.cnil.fr/vos-obligations/decla
rer-a-la-cnil/declaration-videosurveillance/), while a
CCTV camera installed in the UK would have to
follow the UK legal framework and the
recommendation of ICO, the UK Data Protection
Authority (https://ico.org.uk/for-organisations/guide-
to-data-protection/cctv/). Such obligations would be
part of a national common reference created for use
in France and UK respectively. In practice there
MODELSWARD2015-3rdInternationalConferenceonModel-DrivenEngineeringandSoftwareDevelopment
668
could be many socio-contextual and ethical as well
as legal references. While technology references
could be common, it is likely that variations will
exist as some technologies could be more important
in some contexts (e.g accountability features would
be important in Europe).
Further SALT frameworks could also include
proprietary reference information. The position
taken by PARIS is that the success of a SALT
framework depends on the success of a
multidisciplinary practice ecosystem. Several
scenarios are possible. One of them could be the
following: a public authority in a country decides to
build a SALT framework and require surveillance
system designers to use it. Surveillance system
designers would use it as the starting public point,
possibly enriching it with their corporate extended
references. These extensions might not have to be
made public for competitive advantages reasons.
5.3 SALT Framework Usage Challenge
One of the challenges of this whole design process is
to agree on the way a SALT framework would be
used and its value. Many discussions took place in
the PARIS project on this issue with some
stakeholders being concerned that relevant design
decisions would be taken automatically, while other
stakeholders being concerned that maximum
automation would be needed to cope with error
prone aspects. The following was decided in the
PARIS project: interactions between a designer and
SALT frameworks would be based on two
interactions paradigms, the browsing paradigm and
the questionnaire paradigm. Browsing would be
convenient when the designer knows the items it
wishes to browse. For instance, (s)he could look for
a particular set of PETs, AETs or surveillance
technology. Browsing would have less interest in the
early stages of a SALT framework, i.e. when not
many references have been stored, or when
designers are not yet familiar with the design
process. By contrast, Questionnaires would be more
useful at that stage because they would provide
guidance of complex design aspects or just when
designers are not yet familiar. They would become
cumbersome and tedious when used repeatedly, so
they should have several structures, one with
sequential access when users are not familiar with
the content and one with direct access when users
are already familiar with the questionnaire.
Consequently, the structure of a questionnaire is
important, and it must be easy to change. In other
words, knowledge on a questionnaire structure is an
integral part of a SALT framework.
5.4 SALT Framework Management
Challenge
The management of a SALT framework is an
underestimated challenge. This is because the initial
version of a repository is most often created by the
very same researchers and engineers who design the
repository itself. They can do it easily and they focus
on the implementation of the repository not on its
content. There are two problems to tackle. First,
editing and adding content to the SALT framework
must be a mainstream activity based on quality
production tools. Secondly, the governance of the
SALT framework content must be addressed. How
do we decide that some information can be
integrated in a framework? Here are two examples
of problematic governance schemes: an approach
where every contribution is accepted could easily
lead to a vast amount of reference information that
are hard to use; an approach where each reference is
duly validated and certified would create a
bottleneck as the time it would take to create an
entry in a framework could be too long. The PARIS
project has decided to address these challenges by
developing a specific editing tool. The requirements
for such tool are listed in the next section. The
governance of reference context would be based on
transparent peer reviews. For instance, the creation
of a technology entry concerning a PET would be
provided by expert A and reviewed by experts B, C
and D.
6 SALT FRAMEWORK TOOLS
REQUIREMENTS
In order to address such requirements, a number of
use cases were defined (see (Deliverable D2.2,
http://www.paris-project.org/index.php/deliverables)
for some examples). They broadly showed two types
of requirements, those that will be useful to experts
in charge of creating and maintaining a SALT
framework, and those that will be useful to experts
designing privacy preserving surveillance systems.
We consequently need to manage two types of tools:
tools for SALT Framework experts and tools for
surveillance systems designers. We cover them in
sequence.
SALTFrameworkstoTackleSurveillanceandPrivacyConcerns
669
6.1 Requirements for SALT
Framework Experts
The requirements are threefold: reference
management, terminology management and
questionnaire management.
Reference management allows for the creation,
editing, rating and version management of
references, reference structures and viewpoints. If a
SALT framework expert wishes to include a
reference to the European legislation, (s)he includes
a reference in the form of an abstract, a historical
context, and a PDF file containing the text of the
European legislation. (S)He also amends the existing
questionnaires to take into account the new
legislation. The abstract is later amended with an
additional comment highlighting a specific aspect of
the questionnaire. The link from the questionnaire to
the comment is also integrated in the framework. In
order to ensure increasingly better questionnaires,
rating of questionnaires is also supported, at two
levels: by peer experts and by users (i.e. designers of
surveillance systems based on the SALT
framework).
Terminology management allows for the
creation, edition, rating and version management of
taxonomy and terms. It would enable consistent use
in questionnaires. An example of feature could be
equivalence of terms and subclass dependency, i.e. a
questionnaire can use a synonym or a refined term.
A change in a terminology would trigger detection
of changes in questionnaires (and possibly edition of
the questionnaires). It would also trigger version
management, i.e. it would be possible to run the
amended part of the questionnaire only. If a
technology expert creates a taxonomy on
surveillance technologies, (s)he uses the SALT
framework edition tool to store the taxonomy. The
result is subsequently used by further experts to
store entries on surveillance technologies. Some
years later when hundreds of surveillance
technologies have been stored, a new class of
surveillance technology imposes a change in the
taxonomy. The initial expert wishes to change the
taxonomy, without compromising the already
entered entries. For instance socio-contextual,
ethical, legal and technology experts have combined
their knowledge to design one or several
questionnaires. The questionnaires structures are
also stored in the framework. The questionnaires
consist of free text carefully crafted by the experts,
except that surveillance technologies terminology is
based on the taxonomy. The result of changing the
surveillance taxonomy has an impact on the
questionnaire. But the expert is presented with the
options of changing automatically the questionnaire
(when term A is just replaced by term B), or of
being displayed individual occurrences (when term
A is replaced by two other terms).
Questionnaire management allows for the
creation, edition, rating and version management of
questionnaires. A question could be changed, refined
into several questions. The sequence of questions
could be changed, and dependencies between
highlighted could be annotated. A change in a
questionnaire would trigger version management,
i.e. it would be possible to run the amended part of
the questionnaire only. The third example focuses on
the structure of questionnaires: a SALT framework
expert creates a questionnaire which is used by
designers of specific surveillance systems based on
biometric systems in country A. The questionnaire is
used during six months by an initially small number
of designers. Feedback is used by SALT framework
to improve the questionnaire. The sequencing of
some questions are changed. Some questions are
refined into more precise questions. The
questionnaire is then used satisfactorily during a
couple of years until some new regulation
procedures are put in place. The questionnaire is
changed accordingly with the expert determining the
list of previous questions that need not be answered.
A version management capability allows for the
generation of two types of sessions, a new design
session when the new questionnaire is entirely run,
and a redesign session when designed systems using
the previous questionnaire must be verified.
6.2 Requirements for Designers
Here the requirements are twofold: Design
management and Governance management. We
illustrate them through two examples of scenarios
involving surveillance system designers using a
questionnaire in a SALT framework.
Design management allows for the management
of design sessions with design options. Flexible
access, e.g. direct access to questionnaires is
possible (e.g. direct access to a question). Access to
previous designs is also possible to help for reuse. It
also allows for the generation of design
documentation (e.g. a PIA document). The designer
creates a design session and runs the various
questionnaires prepared in the SALT framework. He
selects some parameters providing indication on the
size of the system being designed. He knows that the
entire design lifecycle will take several months.
Consequently he creates a design session that
MODELSWARD2015-3rdInternationalConferenceonModel-DrivenEngineeringandSoftwareDevelopment
670
contains incomplete and sometimes preliminary
answers to questions. Initially he is not familiar with
the questionnaire structure but he is provided with a
high-level structure presentation of the questionnaire
that he can easily match with its system design
process. Later he is familiar with the questionnaire
structure and uses direct access to specific questions.
He can also consider different design options and
weigh their privacy impact, i.e. the questionnaire
session manager can handle several options in
parallel.
Governance management allows the
qualification of answers and the involvement of the
whole design teams (including lawyers, privacy
managers, managers, designers) when some
decisions need more governance. It also allows for
the access to a design dashboard and the creation of
a logbook including in particular accountability
statements. The second example focuses on the use
of the tool to validate and review a designer’s
decision: designers can qualify their answer
according to their design certainty (e.g. designer has
committed to an answer and can justify it, designer
has several design options that is he is analysing,
designer needs further discussions with other
stakeholders in his organisation). He is provided
with a design dashboard allowing him to see the
design status. His design is also accessible by the
corporate members involved in design and its
validation. At the end of his analysis he is left with
two design options which need management review.
Managers access the design session and make a
decision. The rationale for the decision is stored and
kept for traceability and accountability.
6.3 A Tooling Approach based on
Model-Driven Engineering
Techniques
An initial set of tool requirements for SALT
framework experts was defined in (Deliverable
D3.1, http://www.paris-project.org/index.php/deli-
verables). The tool, called SALT framework
management tool (SFMT), relies on the use of
models and model driven engineering techniques
(http://en.wikipedia.org/wiki/Model-driven_enginee-
ring), an approach at the forefront of software
engineering. The main benefit of model driven
engineering is that it allows for the modelling of any
type of knowledge, its representation into suitable
digital form, and the subsequent provision of
associated tools. Here is one example: geographic
maps designed with model driven engineering
techniques would consist of meta-models and
models. Meta-models would describe a map
structure (e.g. roads, lights), include properties (e.g.
one way road, toll, speed limit), and constraints (e.g.
a temporary construction will block the road). The
meta-model would then be instantiated into a model
with a set of data concerning e.g. Belgium. A driver
could then use a navigation tool that gets updated
with dynamic information such as temporary road
constructions, traffic jams and so forth.
SFMT is structured around the need to support
four phases: the SALT knowledge capture phase,
when experts enter in a SALT knowledge repository
references, terminology, questionnaires; the SALT
knowledge analysis and representation, when
specific knowledge on references, terminology and
questionnaires are selected to be represented as
models; the SALT knowledge repository phase, when
the knowledge is stored; the SALT knowledge
application phase, when specific tools are created
for use by surveillance systems designers. Figure 2
depicts the four phases.
Figure 2: SALT Knowledge Process.
6.4 Models Needed to Support SALT
Framework Experts
Many discussions took place in the PARIS project to
identify which type of knowledge needed to be
modelled. It was agreed that modelling techniques
should not be used to systematically capture the
semantics of referenced information (e.g. a legal
text). Rather they would be useful to capture the
semantics of structures for multidisciplinary
practice. Consequently, models have been identified
for the following needs: Concerning reference
management, reference viewpoint structures will
SALTFrameworkstoTackleSurveillanceandPrivacyConcerns
671
provide viewpoint models for each type of
stakeholder (e.g. the technology viewpoint for a
legal reference is an adapted abstract associated with
the reference). Concerning terminology
management, taxonomies will provide dependency
models that can be used to create flexible
questionnaires. Concerning questionnaire
management, questionnaires’ structure will provide
question grouping and sequencing capability.
Relationships between references, taxonomy and
questionnaires will provide consistency in a
questionnaire with references and terminology.
PARIS also anticipates that other types of
knowledge could be modelled, for instance related to
technical considerations leading to decisions on
PETs. Those additional models, and associated
engineering tools can easily be added in the future to
a SALT framework repository already based on
models.
6.5 Towards Models Checking
One of the potential benefit of models to represent
SALT framework knowledge is model checking. For
instance SALT modelled references would be
compliant to a specific format with metadata added
concerning authors, date, or purpose. Constraints
written with a specific language called Object
Constraint Language (OCL) could then be added, for
instance on data retention for a video camera.
PARIS has developed a specific tool called
PAERIS (Deliverable D4.3, http://paris-
project.org/index.php/deliverables) which checks
that constraints are valid in a developed surveillance
system model (figure 3)
Figure 3: Model Checking for Surveillance System
Design.
7 CONCLUSIONS
This paper has presented the approach proposed in
the PARIS project to integrate surveillance and
privacy issues into a privacy-by-design process,
based on the use of the concept of SALT
frameworks and an associated SALT Framework
Management Tool (SFMT). The PARIS project is
currently implementing a SFMT and applying it to
two use cases, a video surveillance lifecycle
management use case (Deliverable D5.1,
http://www.paris-project.org/index.php/deliverables)
and a biometric surveillance use case (Deliverable
D6.1, http://www.paris-project.org/index.php/
deliverables). It is also implementing a prototype
model checking verification tool, PAERIS.
This article has shown that in the field of
surveillance, the PbD process should give specific
weight to multi-stakeholders empowerment, to
mechanisms that enable to carry out the
proportionality assessment (such as concerted
impact assessment) and finally to accountability and
transparency tools. The SALT framework developed
within the PARIS project builds on these three
pillars. It has also been shown that while the PbD
process can rely on general principles, it is
paramount to specify these principles to the
dedicated environment to which they should be
applied. Finally, it is too soon to assess the need for
specific features or standards, as this will depend on
the needs of the SALT framework ecosystem.
ACKNOWLEDGEMENTS
This paper was made possible thanks to the funding
and work performed for the PARIS project (PrivAcy
pReserving Infrastructure for Surveillance). This
project has received funding from the European
Union’s Seventh Framework Programme for
research, technological development and
demonstration under grant agreement no 312504.
We would also like to thank Ann Cavoukian who
suggested the use of the term contextual in the
SALT acronym.
REFERENCES
http://en.wikipedia.org/wiki/Edward_Snowden.
See for an overview, Electronic Frontier Foundation,
EFF's Cheat Sheet to Congress' NSA Spying Bills,
https://www.eff.org/fr/deeplinks/2013/08/effs-cheat,
11 September 2013.
CJEU, Joined Cases C-293/12 and C-594/12, Digital
Rights Ireland Ltd (C-293/12) v Minister for
Communications, Marine and Natural Resources and
Others and Kärntner Landesregierung (C-594/12) and
MODELSWARD2015-3rdInternationalConferenceonModel-DrivenEngineeringandSoftwareDevelopment
672
Others. Judgment (Grand Chamber) of 8 April 2014,
§37, http://curia.europa.eu/juris/liste.jsf?num=C-
293/12.
Reform of the Data Protection Package,
http://ec.europa.eu/justice/data-
protection/review/index_en.htm.
Privacy-by-Design.
http://www.ipc.on.ca/english/Privacy/Introduction-to-
PbD/
Wright, David; de Hert, Paul (Eds.). Privacy Impact
Assessment. Series: Law, Governance and Technology
Series, Vol. 6, Springer. 2012.
PIAF: Privacy Impact Assessment Framework.
http://www.piafproject.eu.
S.Spiekermann, L.Cranor. Privacy Engineering. IEEE
Transactions on Software Engineering, Vol. 35, Nr. 1,
January/February 2009, pp. 67-82.
S. F. Gürses, C. Troncoso, and C. Diaz. Engineering
Privacy-by-Design. Computers, Privacy & Data
Protection, 2011.
A.Kung, J.Freytag, F.Kargl. Privacy-by-design in ITS
applications. 2nd IEEE International Workshop on
Data Security and Privacy in wireless Networks, June
20, 2011, Lucca, Italy.
http://pripare.eu/
ISO/IEC JTC1/SC27: http://www.jtc1sc27.din.de.
OASIS. Organization for the Advancement of Structured
Information. https://www.oasis-open.org/
OASIS Privacy Management Reference Model (PMRM)
Technical Committee. https://www.oasis-
open.org/committees/pmrm/charter.php.
OASIS Privacy by Design Documentation for Software
Engineers (PbD-SE) Technical Committee.
https://www.oasis-open.org/committees/pbd-
se/charter.php.
http://ec.europa.eu/DocsRoom/documents/5290.
PARIS: PrivAcy pReserving Infrastructure for
Surveillance. http://www.paris-project.org/
CNIL methodology for privacy risk management.
http://www.cnil.fr/fileadmin/documents/en/CNIL-
ManagingPrivacyRisks-Methodology.pdf.
Surveillance, fighting crime and violence. Deliverable
D1.1, IRISS Project. February 2012.
http://irissproject.eu.
Antonio Kung. PEARs: Privacy Enhancing ARchitectures.
Annual Privacy Forum, May 21-22, 2014, Athens,
Greece. Proceedings APF14 "Privacy Technologies
and Policy". Springer Verlag.
Deliverable D2.2 Structure and Dynamics of SALT
Frameworks http://www.paris-project.org/index.php/
deliverables.
http://en.wikipedia.org/wiki/Architectural_pattern.
http://www.cnil.fr/vos-obligations/declarer-a-la-
cnil/declaration-videosurveillance/
https://ico.org.uk/for-organisations/guide-to-data-
protection/cctv/
Deliverable D3.1 SALT Framework Management Tool
Requirements http://www.paris-project.org/index.php/
deliverables.
http://en.wikipedia.org/wiki/Model-driven_engineering.
Deliverable D5.1 Video Surveillance Lifecycle
Management Use Case Description http://www.paris-
project.org/index.php/deliverables.
Deliverable D6.1 Biometrics Use Case Description
http://www.paris-project.org/index.php/deliverables.
Deliverable D4.3 SALT Compliant Processes Guidelines
for Use Cases - http://paris-project.org/
index.php/deliverables.
SALTFrameworkstoTackleSurveillanceandPrivacyConcerns
673