FoT-Rules: A Semantic Rule-based Approach for Smart Spaces Through Fog of Things

Tracking #: 2072-3285

Cleber Santana
Brenno Alencar
Ernando Batista
Cassio Prazeres

Responsible editor: 
Guest Editors Sensors Observations 2018

Submission type: 
Full Paper
Smart Spaces or Smart Environments are related with ubiquitous computing in the sense that sensors, actuators, and others computational elements should be embedded seamlessly in the everyday objects. In the Internet of Things (IoT), Smart Spaces will enable environments to adapt according to people (users) needs by using smart and connected objects. However, to turn the IoT view into a reality, the users should know about technical details of such objects, which is not a trivial task for most ordinary users. Therefore, this article presents FoT-Rules, a proposal for the construction of semantic rules aiming to create Smart Spaces through Fog of Things, which is a paradigm for Fog Computing in the Internet of Things. FoT-Rules is designed to enable ordinary users to create semantic rules in the Event-Condition-Action standard (ECA) and to take actions according to the environment at the edge (Fog) of the network. In this work, we present a scenario, where the user can create semantic rules in the ECA standard, and then FoT-Rules perform the following functionalities: (i) obtain the semantic model that contains information related to an IoT device; (ii) execute a semantic reasoner over the semantic model according to the rule created by the user; (iii) provide a semantic observer that is responsible for observing changes in IoT devices; and (iv) in case the rule created by the user is activated, an action is taken for an IoT device. Finally, we performed three types of evaluation on our FoT-Rules proposal: reliability, efficiency and usability.
Full PDF Version: 

Major Revision

Solicited Reviews:
Click to Expand/Collapse
Review #1
By Payam Barnaghi submitted on 26/Feb/2019
Minor Revision
Review Comment:

This paper describes a rule based system to describe event-condition-action statements and check the consistency of the rules in IoT environments. The authors have described a mobile application for creating and running the rules and have also described the semantic modelling and rule and consistency checking for the designed framework.

(1) originality: the work proposes an interesting application for the IoT systems; the idea of using higher-level abstractions and semantic rules for designing and mangling IoT applications is a very good approach. However, the presented rules seem to be based on simple ideas and the novelty of the solution beyond the rule creation and consistency checking is not clear. Other aspects such as adaptability to changes, fault tolerance, imprecise conditions, complex scenarios, etc. could be also considered in the research.

(2) significance of the results: the work presents a relatively interesting end-to-end application; however, the scalability and applicability of the work to complex and real-world scenarios are not discussed. It is not clear how the system will guide a user to define/identify conditions for each data stream/source; how the processing and analysis will be performed (or if a their-part application will do this, perhaps an example could be provided). The paper would have benefited from a real-word use-case scenario.

and (3) quality of writing: overall well written; however, there is not link to the code and/or an online demo or the semantic rules that are created.

Overall this is a good paper. It presents an end-to-end solution for IoT applications using semantic rules and task planning based on event-condition -action statements; however, the novelties of the work seem to be limited.

Review #2
By Antonio Piccinno submitted on 28/Feb/2019
Major Revision
Review Comment:

The paper proposes FoT-Rules, an approach based on semantic rules built into the ECA format to support end-users on the creation of Smart Spaces at the edge of the network. FoT-Rules is an extension the Fog of Things (FoT) paradigm proposed by other authors for designing and implementing Fog Computing platforms for the Internet of Things.

The overall approach presented in the paper is very interesting and original, but my feeling about the manuscript is that it contains many concepts and elements, often presented in a poor and not coherent way that makes the reading very difficult and confused.
For example, the title promises an approach (FoT-Rules, which is the focus of the paper) that becomes a scenario in the abstract “that perform functionalities”, then an approach and then again a usage scenario. FoT Rules at some points is also referred as modules and then seems to be a system.
It is very difficult to understand the real flow of the paper. For example, the authors use a lot of remind to other sections (before or after) that should be avoided (see for example Section 3.1 with many remind to Section 4). A running example throughout the paper would help the reading for sure.

Furthermore, the manuscript addresses semantic rules in ECA standard for a single device but also refer to smart space. What about the orchestration of the different devices? The authors should also consider this aspect.

Three evaluation are reported, but they are lacking and, sometimes, very trivial and should be better sustained, not only through tables with data.

I’m convinced about the validity of the approach, but the manuscript cannot be accepted in the current status. Many errors and typos should be corrected.

In the following my detailed comments.

Works in Section 2.1 should be related to the proposed approach, not only listed. For example the work from Kaed et al. reports about orchestration, why the author did not consider this aspect?

Section 2.2 contains a list of characteristic of a semantic approach not yet presented in the text: Table 1 contains the comparison with other works by simply a checkmark. A deeper analysis and discussion should be provided. I suggest to merge 2.1 and 2.2 and provide a more harmonious related work section.

Examples of SOFT-IoT components listed in Section 3.1 would help the reader to understand the underlying system for FoT-Rules.

Figure 1 is a complex one and is full of details that are missing in the textual description. An overall architecture of the system presented earlier is needed. In the text the author speak about module (e.g., Semantic Model and Semantic Reasoner), but they are not in any architecture.

The example reported at the end of Section 3.2 does not explain how the reasoner works. What the sense?
Furthermore, it is not clear the example in Listing 2 to what is referred to.

OSGi (introduced in Section 4.1) should be presented in detail and FoT-Gateways and FoT-Servers identified in the figure (or in the possible architecture). It is not clear which one are the FoT-Rules modules” in Figure 3.

Figure 4 shows a screen shot of a mobile application, not the mobile application as stated in section 4.2.2. Its functionalities should be described and discussed. Furthermore, the authors state: “the user chooses from a list the sensor for which s/he wishes to create rules and informs the appropriate values FoT-Servers”. Where and how?

The design of the evaluation study should be first presented. Results need to be discussed (and not only reported in a table), and “threats to validity” added. Why Table 3 and 4? Why such scenes?

Finally, in Section 6. The authors speak about instantiating FoT-Rules in other domains and in other scenarios, but it is not clear which domain they refer to. They means out of smart spaces?

Conclusion are very poor and trivial.

Some typos
Page 8: “FoT-Rules modules is”
Fig. 3: “Micoservices”
Section 6: “studied as a menas to”

Review #3
Anonymous submitted on 25/Mar/2019
Major Revision
Review Comment:

The paper presents a rule-based approach to create ECA rules via an interface that supports end-users for constructing smart spaces at the edge of the network. By executing a reasoner over the semantic model of an IoT device and the created rules by the user, an action is taken for the device.

I believe it is important for the IoT community to devise approaches towards the ordinary users as a way to showcase the usability of their technologies, test and compare their performance, and push them forward in order to address new challenges. The topic of the paper is appropriate to the journal and the promise (from the title and abstract) that a smart space is established by an ordinary user through an interface that generates rules is met with enthusiasm. The related work covers a good amount of relevant studies and provides a neat comparison table which clearly shows the novelty of the proposed framework. However, the submission in its current form somehow does not tick all the boxes to be up to the overall standard of SWJ.

First, I think the main contribution of the paper is not well framed and hard to point out in the introduction. The line of argumentation gets difficult to follow due to many pros and cons provided while describing the chronological development of the state-of-the-art. The paragraph describing the benefits of a semantic-based representation model is not self explanatory (e.g., what is reactivity and what is information hiding?) and it feels like the beginning of a new part rather than an organic part of the section. I would expect to see a better organized introduction which develops consistently towards a clearly and boldly stated contribution.

Second, the approach is not represented in a formal way but it is rather described gradually with the running scenario in Section 3. As a suggestion, the core technical contribution and the scenario can be separated. Moreover, the running example for creating rules for the two sensors is presented with a fair amount of redundancy (e.g., same paragraph appears twice on the page 8). Given that the concepts do not seem too complicated, this separation should be rather easy to achieve.

Concerning the evaluation, another more complex scenario with more rules (and maybe with the rules that have side effects on each other) than the running scenario could be presented only for the reliability evaluation. In this way, the approach would be tested in a more demanding setting. A link in the evaluation section for accessing the implementations of the reliability and the efficiency evaluations (5.1 and 5.2) would also strengthen the section so that any reader could potentially reproduce these results. Regarding the usability evaluation (5.3), as valuable as the overall positive results are, including some commentary by the participants who “totally disagree” with some of the statements may provide the readers with a qualitative summary along with the quantitative charts.

Please, find in the following some of the typos and other remarks:
- All the figures are bitmap images and therefore especially some of them (e.g. Fig. 3) does not scale well. I strongly suggest the authors to rework all the figures as vector images and to provide a visually appealing and consistent style.
- Page 5, subsection 3.1: SOTF-IoT → SOFT-IoT
- Page 7, subsection 3.2: an FoT-Gateway → a FoT-Gateway
- Page 8, figure 2: The flowchart has an unusual styling, e.g. the start node and end node shapes are uncommon, Comic Sans font looks inappropriate. The arrow from (1) to (2) is in the wrong direction.
- Page 8, listing 2, Line 3: Is the closing parenthesis a typo?
- Page 9, subsection 4.1: “Such server” → “Such a server”?
- Page 9, subsection 4.1: “…mobile application and do not...” → “…mobile application and does not...”
- Page 9, subsection 4.2.1: In the paragraph starting with “Listing 3 shows some...”, the line number references may be wrong. The dot is missing at the end of the sentence.
- Page 9, subsection 4.2.2: “In Rule A, it is inferred that temperatures higher than 40 C are considered high (highTemperature true).” may be wrong. I think “(?a highTemperature true)” in Listing 4 sets the highTemperature attribute of the device ?a as ‘true’.
- Page 11: Listing 7 can be positioned at the top of the page.
- Page 12, subsection 5.1: Abnormal conditions and errors can be further explained.
- Page 12, subsection 5.2: Regarding the sentence “...the test time result for two samples…as illustrated in Section 4.”, ‘samples’ are not mentioned before, therefore it’s not easy to understand that a ‘sample’ refers to a ‘problem input’. Moreover what “illustrated in Section 4” is also not clear.
- Page 12, subsection 5.2: Does “...with 10 steps” mean “...with 10 repetitions”?
- Page 13: Line spacing is irregular.
- Page 14, table 3: (in the repetition column) 58 → 8
- Page 14, table 5: The statement number column is missing (regarding the connection to Fig.7-10).
- Page 15, figure 6: The figure header is redundant and can be presented in the caption.
- Page 15, section 6: menas → means
- References are not presented in a uniform style. Also please check editors, cities, abbreviated names vs extended names of journals and conferences, and so on.