UMLsec4Edge Extending UMLsec to model data protection related requirements in edge computing Sven Smolka1 Jan Laufer1 Zolt anAdam Mann2 Klaus Pohl1

2025-05-06 0 0 332.67KB 8 页 10玖币

侵权投诉

UMLsec4Edge: Extending UMLsec to model data

protection related requirements in edge computing

Sven Smolka1, Jan Laufer1, Zolt´

an ´

Ad´

am Mann2, Klaus Pohl1

1paluno – The Ruhr Institute for Software Technology, University of Duisburg-Essen, Essen, Germany

2Complex Cyber Infrastructure (CCI) group, University of Amsterdam, Amsterdam, The Netherlands

Abstract—Edge computing enables the processing of data –

frequently personal data – at the edge of the network. For

personal data, legislation such as the European General Data

Protection Regulation requires data protection by design. Hence,

data protection has to be accounted for in the design of edge

computing systems whenever personal data is involved. This

leads to speciﬁc requirements for modeling the architecture of

edge computing systems, e.g., representation of data and network

properties.

To the best of our knowledge, no existing modeling language

fulﬁls all these requirements. In our previous work we showed

that the commonly used UML proﬁle UMLsec fulﬁls some of

these requirements, and can thus serve as a starting point.

The aim of this paper is to create a modeling language which

meets all requirements concerning the design of the architecture

of edge computing systems accounting for data protection. Thus,

we extend UMLsec to satisfy all requirements. We call the

resulting UML proﬁle UMLsec4Edge. We follow a systematic

approach to develop UMLsec4Edge. We apply UMLsec4Edge to

real-world use cases from different domains, and create appro-

priate deployment diagrams and class diagrams. These diagrams

show UMLsec4Edge is capable of meeting the requirements.

Index Terms—edge computing, data protection, UMLsec, fog

computing

I. INTRODUCTION

Motivation: Edge computing enables real-time data pro-

cessing with low network communication latency and satisfac-

tory quality of service at the same time. Due to these positive

characteristics, edge computing systems become increasingly

popular [1]. So-called edge nodes can perform tasks that most

end devices (e.g. IoT devices) are not capable of due to

insufﬁcient computing power [2]. Edge nodes can also pre-

process data, for example to reduce the amount of data that is

sent to the cloud for further processing [1]. However, process-

ing (personal) data at the edge of the network leads to new

challenges. For example, edge nodes may use different types

of communication (e.g., 5G or WLAN) which offer different

levels of data protection. Regulations such as the General

Data Protection Regulation (GDPR) of the European Union

[3] prescribe the protection of personal data. In particular, the

GDPR stipulates the need for data protection by design. To

ensure data protection by design in edge computing systems,

requirements for modeling the architecture of edge computing

systems emerge [4] which can be sorted into categories: “net-

work properties”, “devices”, “actors”, and “data properties”.

Work partially funded by the European Union’s Horizon 2020 research and

innovation programme under grant agreement no. 871525 (FogProtect). Useful

discussions with project partners are gratefully acknowledged.

Problem statement & Aim of the paper: Existing model-

ing languages used in system design focus either on security

(e.g. [5], [6]), or on privacy (e.g. [7], [8]), or none of the

above. Although the GDPR requires appropriate security under

Article 32, the deﬁnition of data protection goes beyond that

of security and privacy. Existing modeling languages cannot

capture all aspects of edge computing systems related to data

protection, for example, the different levels of data protection

stemming from different types of communication technologies

[4]. To the best of our knowledge, there is no modeling

language which covers all requirements on the modeling of

the architecture of data-protection-compliant edge computing

systems. We identiﬁed UMLsec (an extension of the well-

known modeling language UML) as a promising starting

point to model data protection concerns in edge computing

systems, because UMLsec is capable of modeling information

security aspects for software systems. UMLsec addresses the

“data properties” requirement category as well as parts of the

categories “network properties” and “devices”.

The aim of this paper is to create a modeling language which

satisﬁes the identiﬁed requirements concerning the design of

the architecture of edge computing systems, accounting for

data protection.

Contribution & Approach: To achieve this goal, this paper

provides the following contributions.

•A new modeling language called UMLsec4Edge that

supports the modeling of data-protection-compliant edge

computing systems. UMLsec4Edge is an extension of

UMLsec, focusing on deployment and class diagrams. We

focus on these diagram types because the architecture of

systems and dependencies between system components

have great impact on data protection by design.

•A formalization of UMLsec4Edge as well as an UML

proﬁle created with the modeling platform Papyrus.

•Deployment and class diagrams of three real-world use

cases which employ edge computing in diverse appli-

cation domains. The use cases originate from the EU-

funded research project FogProtect [9]. The diagrams

conﬁrm that UMLsec4Edge meets the requirements by

addressing the shortcomings of UMLsec.

The creation of UMLsec4Edge follows a systematic approach

based on the work of Lagarde et al. [10], including a literature

search focusing on UML proﬁles covering security, privacy, or

data protection to assess alternative extension options.

arXiv:2210.09358v1 [cs.SE] 17 Oct 2022

Outline: Sec. II deﬁnes terminology and introduces the

requirements for modeling the architecture of edge com-

puting systems accounting for data protection. Sec. III de-

scribes our systematic development approach and as a re-

sult the UMLsec4Edge proﬁle. Sec. IV shows extracts from

UMLsec4Edge diagrams and discusses threats to validity. Sec.

V examines related work, while Sec. VI concludes the paper.

Additional material, including the full UMLsec4Edge proﬁle,

complete diagrams, and further information on the systematic

approach and literature review, can be found online1.

II. REQUIREMENTS TO MODEL DATA PROTECTION IN EDGE

COMPUTING SYSTEMS

A. Data protection in edge computing systems

The term data protection refers to the protection of per-

sonal data. Regulations such as the GDPR [3] prescribe this

protection. The GDPR requires the enforcement of technical

and organizational measures to prevent so called personal

data breaches. Personal data breaches occur, e.g., whenever

an unauthorized actor accesses personal data. Any system, and

therefore also an edge computing system, which processes per-

sonal data must ensure the absence of personal data breaches.

Ensuring data protection in edge computing systems faces

new challenges compared to more established computing

paradigms like cloud computing. End devices and edge nodes

could differ in hardware conﬁguration, preventing the imple-

mentation of uniform data protection mechanisms such as

hardware enclaves across all devices. Moreover, end devices

and edge nodes can be deployed almost anywhere. Thus, they

may not be protected by sufﬁcient physical security measures,

so there is a threat of attackers physically compromise them.

Since the GDPR requires systems processing personal data

to ensure data protection by design, such data protection

challenges need to be considered when developing an edge

computing system, already starting with the architectural de-

sign of the system. In our previous work [4], we identiﬁed four

data-protection-related requirements on modeling languages

for modeling the architecture of edge computing system:

(R-1) Network properties: It must be possible to model

different communication types between devices in an edge

computing system as well as possible threats posed by them.

(R-2) Devices: It must be possible to model different device

types in an edge computing system as well as threats posed

by their use.

(R-3) Actors: It must be possible to model actors within an

edge computing system, as well as their trust in each other,

their relationship to data, and their data-speciﬁc roles.

(R-4) Data properties: It must be possible to model data

protection requirements specifying whether data must not be

disclosed, manipulated, or deleted.

B. Modeling edge computing architectures with UMLsec

In our previous work [4], we analyzed to what extent

UMLsec [5] supports modeling data protection requirements

1See https://git.uni-due.de/fogprotect/umlsec4edge

and threats to data protection during the design of edge

computing systems. We concluded that UMLsec provides a

reasonable basis for satisfying the identiﬁed requirements:

(R-1): UMLsec introduces stereotypes such as <<wire>>

or <<LAN>>, which can be used to assign a connection type to

a communication path between nodes in deployment diagrams.

In addition, UMLsec introduces the adversary model, which

represents the threat of unauthorized reading, insertion, and

deletion of data during data exchange over a communication

path of a certain connection type. UMLsec is limited in

having only stereotypes representing wired connection types.

In edge computing systems, however, data exchange between

nodes often takes place wirelessly. Consequently, UMLsec

only partially addresses (R-1).

(R-2): UMLsec allows assigning device types to nodes by

stereotypes like <<POS device>>. The adversary model

then allows the representation of the threat of unauthorized

physical access to these types of devices. However, there are

no stereotypes for device types common in edge computing

systems. Accordingly, the threat of unauthorized physical

access to them cannot be modeled in the adversary model. In

addition, UMLsec cannot model threats occurring when data

is exchanged between components placed on the same node.

Overall, UMLsec fulﬁlls a part of (R-2).

(R-3): UMLsec has no stereotypes or tags to model actors

within an edge computing system as well as their trust in

each other, their relationship to the data, and their data-speciﬁc

roles. Accordingly, UMLsec does not address (R-3).

(R-4): UMLsec introduces stereotypes and tags to model

the security requirements of data during data exchange in

terms of conﬁdentiality, integrity, and availability. For ex-

ample, UMLsec introduces the stereotype <<secrecy>>,

which can be attached to dependencies between nodes and

components in a deployment diagram to deﬁne the security

requirement of disallowing data to be read by an attacker

during data exchange. In combination with the adversary

model (which represents the threats of data transmission over

a communication channel with a speciﬁc connection type), it

is possible to evaluate whether security objectives are met by

the system design. Thus, UMLsec addresses (R-4).

Table I summarizes the restrictions that make UMLsec not

fulﬁll the requirements. The table also lists the stereotypes and

tags of our UMLsec4Edge proﬁle (presented in the following

section) leading to the fulﬁllment of the requirements.

III. UMLSEC4EDGE

A. Systematic Approach towards UMLsec4Edge

To create our UMLsec4Edge proﬁle that satisﬁes all the

requirements, we conduct a systematic extension of UMLsec.

Since UMLsec already partially addresses (R-1) and (R-2)

in deployment diagrams, it is reasonable to extend UMLsec

with respect to deployment diagrams to fully address the

requirements. As UMLsec does not address (R-3), we extend

UMLsec with respect to class diagrams, since the data of a

system is often modeled in class diagrams. Fig. 1 shows our

systematic approach of extending UMLsec. We ﬁrst create

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

UMLsec4Edge:ExtendingUMLsectomodeldataprotectionrelatedrequirementsinedgecomputingSvenSmolka1,JanLaufer1,Zolt´an´Ad´amMann2,KlausPohl11palunoTheRuhrInstituteforSoftwareTechnology,UniversityofDuisburg-Essen,Essen,Germany2ComplexCyberInfrastructure(CCI)group,UniversityofAmsterdam,Amsterdam,TheNetherl...

展开>> 收起<<

UMLsec4Edge Extending UMLsec to model data protection related requirements in edge computing Sven Smolka1 Jan Laufer1 Zolt anAdam Mann2 Klaus Pohl1.pdf

共8页,预览2页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

UMLsec4Edge Extending UMLsec to model data protection related requirements in edge computing Sven Smolka1 Jan Laufer1 Zolt anAdam Mann2 Klaus Pohl1

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: