Trustworthiness Model Representation
attributes that contribute to an overall trust measure. Hence it is essential to define a generic model to compute trust. The trustworthiness of a solution involves
● Computing a Trust Score for each component
● Computing an overall Trust Score that is a combination of the Trust Scores of each component involved in providing a solution.
The model used for computing a trust measure must be generic enough in order to provide a Trust Score for various applications. The IoT Policy Framework defined by EU commission calls for a similar hierarchical approach to establishing trust. 10
The attributes involved in computing the Trust Score of a component can be classified into two broad classes:
A. Quantitative attributes: Such parameters can be measured numerically. Examples include: a. Temperature values measured by a sensor. b. Date, time and size of a file installed on the system. c. A fingerprint – like a secured hash of a file installed on a system.
B. Qualitative attributes: Such characteristics are assigned a unique label from a set of valid labels( i. e., classification). Typically, such attributes are assigned a string value like a label assigned to an object in an image observed by a camera. Examples could include: a. Vehicle type identification such as truck, forklift or unknown vehicle. A factory could use this kind of qualitative parameter to establish trust for delivery trucks. It would ensure that only authorized vehicle types are docked at the loading stations. b. Part type at an assembly work station. A factory could use this type of parameter to ensure that the correct components are used in an assembly operation. It ensures the trustworthiness and quality of the final product. c. Ingredient name / label at a mixing station. The food industry could use such an attribute to establish trust in the final product, providing for ingredient safety and product quality.
The attributes are further grouped into logical sets. As an example, in a manufacturing solution, the attributes of a component within a manufacturing solution can be grouped into the following logical sets:
● System hardware attributes
● System software attributes
● Application software attributes
● System access attributes
● System temperature attributes
● System vibration attributes
● System configuration attributes
● Observed pattern attributes
Each attribute can be assigned a different weight based on how it affects the overall
10
H2020 – CREATE-IoT Project, https:// european-iot-pilots. eu / wp-content / uploads / 2017 / 10 / D05 _ 01 _ WP05 _ H2020 _ CREATE- IoT _ Final. pdf, 2017.
September 2018- 32-