Skip to main content

Secure Data Processing in the Cloud

Periodic Reporting for period 2 - RESTASSURED (Secure Data Processing in the Cloud)

Período documentado: 2018-07-01 hasta 2019-12-31

"Secure cloud computing is key for business success and end-user adoption. of cloud services, and thus essential to stimulate the growth of the European Digital Single Market. Yet for many business sectors, the idea of hosting sensitive business or personal data on a public cloud raises concerns over the security and privacy of the data; while encryption techniques can protect the transfer of data to and from the cloud (""data-in-motion"") as well as data stored on the cloud (""data-at-rest""), operating on this data requires decryption, leaving data-in-use in computer memory exposed to security breaches. Additionally, the EU's General Data Protection Regulation which became effective on May 25, 2018, brings additional requirements to data retention within the cloud.

RestAssured's goal is to enable the free and seamless movement of data within the EU, while assuring conformance to data protection regulations, as well as offering data security and privacy across the whole life cycle of the data."
WP3: Architecture, Platform and Methodology
WP3 provided the overall project architecture, as well as a physical hardware test bed for code development and integration.

WP 4 Secure Cloud Data Processing and Execution Environment

This WP examined establishing hardware enclaves using hardware with Intel’s SGX technology and
identified ease of use challenges and gaps that prevented adoption of the technology in a Cloud environment. Micro-benchmarking was performed. Additionally, IBM collaborated with the Berkeley RISE lab on the development of Opaque
- an implementation of Apache Spark SQL on top of SGX hardware.

In the second half of the project, this task examined using AMD SEV technology for secure hardware enclaves.

This WP led the development of what became the official standard for Modular Parquet Encryption, which has been accepted by the Apache community.
Additionally, IBM developed a reference version of code which runs Parquet encryption integrated into the hadoop environment.

Advanced features such as the prevention of tampering with encrypted files were also developed. IBM incorporated
this work in its Cloud-based Analytics Engine offering.

WP5: Run-Time Data Protection Assurance
WP 5 delivers novel monitoring and adaptation solutions for detecting
and mitigating violations of data protection policies in the cloud. In the first half of the project, the focus
was on the detection of data protection violations, and in the second half of the project,on adaptation.

Task 5.1: Runtime Engine for Detecting Data Protection Policy Violations
In Task 5.1 UDE devised a novel model-based approach for detecting cloud configurations with an unacceptably
high risk of data protection violation.

Task 5.2: Runtime Engine for Restoring Data Protection Policy Compliance
We devised an approach for data-protection-aware adaptations in
cloud systems. In this approach, the run-time model of the cloud system is continually analyzed using the
methods from Task 5.1 to detect if the cloud configuration is associated with unacceptably high data protection
risks.

Task 5.3: Models@Runtime as a “Shared Knowledge Base”
A first, a proprietary meta-model for data protection in the cloud was introduced in D5.1. Later, the
meta-model was re-engineered to make it conform to the TOSCA standard.

WP6 Decentralized Data Lifecycle Management
Task 6.1 Methodology for Data Lifecycle Management
In accordance with the work plan, the work in WP6, which was led by Thales, focused on the methodology
for Data Lifecycle Management.

In Period 2, the progress of this task mainly focused on the revision of the secure data lifecycle phases
and processes. The runtime phases that were not covered in period 1 were also developed by integrating the context-based
approach, namely the compliance checking and the change management and adaptation process.

WP 7 Engineering for Run-Time Data Protection Summary
Task 7.1 Security and Privacy by Design Methodology
In the first half of the project, a methodology was devised for incorporating risk assessment into an overall
security and privacy by design approach.

Task 7.2 Security and Privacy Threat Identification Tools
Models were developed to capture the structure of cloud based systems at the required levels. At the
highest level, we have extended the CSAP approach from UDE by developing patterns suited to cloud based
systems. The enhanced SSM tool is now the subject of a commercialisation activity supported by the UK government.

Task 7.3 Security and Privacy Threat Mitigation Tools
During the first half of the project, the main focus was to improve the coverage of
basic cyber security threats and the commonly used countermeasures such as firewall user authentication
and access control, firewall restrictions, software patching and security testing, etc. In the second period,
these threat mitigation models were extended to capture measures introduced by the RestAssured project,
such as the use of secure enclaves.

WP 8 Use Cases and end-user validation
Task 8.1 Validation planning, analysis and evaluation

Implementation and Execution of “Self-directed Social Care” Use Case
In the first half of the project OCC produced SCANT (Social Care Analysis of Needs Tool) to assist local
authorities in identifying unmet social care needs, whilst also preserving the privacy of the potentially
vulnerable Ami users (volunteers and clients).

Implementation and Execution of “Pay as You Go Insurance” Use Case
In the first half of the project, Adaptant implemented a simple Pay-As-You-Drive system on top of the
RestAssured v1 architecture.

In the second half of the project, this use case was extended to look at the introduction of further-restricted
sensitive data (in this case, biometric data) in which the handling of the data is subject to varying requirements
as the vehicle moves between countries.

WP 9 Impact
This work package was responsible for the communication and dissemination of the project results and
establishing a strong presence in the research community through both offline and online channels.

The acceptance of 30 papers for publication in scientific journals and conferences plus several publications
for the general audience in both magazines and online blogs further highlights the major focus for
dissemination activities in this area.

Additional dissemination was achieved through attendance of workshops and collaboration meetings such
as the RestAssured representation in the DPSP (Data Protection, Security and Privacy in cloud) cluster.
Furthermore the RestAssured project coordinated a cluster of projects to complete the CDB
(Common Dissemination Booster).

This work package resulted in the creation and subsequent dissemination of the RestAssured Handbook (D9.6 Final
RestAssured Handbook 1) as part of the project dissemination strategy.
Technical conclusions
RestAssured progressed the SotA both through its scientific publications, and through its technical work.

The work on Parquet encryption standard was started in mid-2017 in RestAssured work as a result of IBM’s
work on SGX and Opaque Encryption of Parquet data followed as a result of an understanding of the requirements of the RestAssured use cases.
The adoption of the Parquet encryption format as a standard by the Apache Parquet community is a major project achievement.