IoT Helper : A Lightweight and Extensible Framework for Fast-Prototyping IoT Architectures
Round 1
Reviewer 1 Report
The proposed work is well written and presented. However, it presents an IIoT framework based on well-known protocols and standards. The work does not compare the proposed solution with other existing works in the IIoT field or the possibility of using protocols different than those selected, i.e. CoAP, missing an introduction on current IoT scenario. The adopted architecture does not take into account the evolution of networks from an Edge-Fog Computing and 5G point of view, despite mentioning them. I would invite the authors to consider widely used frameworks in the IoT field such as oneM2M, OMA-LwM2M and firmware for embedded systems such as Tasmota and Espruino that can be used on more advanced and cheaper hardware than Arduino boards. The core of the work developed by authors (Configuration Module and Generic firmware) could be dealt with in more detail, giving greater prominence to what has been implemented and highlighting the contribution and innovation with respect others firmware.
Author Response
We would like to thank the anonymous reviewers for their comments, which helped us to improve the paper. Following are our answers to specific points raised in this review.
Comment:
The work does not compare the proposed solution with other existing works in the IIoT field or the possibility of using protocols different than those selected, i.e. CoAP, missing an introduction on current IoT scenario.
The adopted architecture does not take into account the evolution of networks from an Edge-Fog Computing and 5G point of view, despite mentioning them.
I would invite the authors to consider widely used frameworks in the IoT field such as oneM2M, OMA-LwM2M and firmware for embedded systems such as Tasmota and Espruino that can be used on more advanced and cheaper hardware than Arduino boards.
Answer:
Thank for this comment, which gave us the opportunity to greatly improve the related work section. We have completely reworked the section, and now separately discuss different hardware platforms for embedded systems, different protocols, different architectures.
In addition, we have added a comparison to some commercial platforms.
Comment:
The core of the work developed by authors (Configuration Module and Generic firmware) could be dealt with in more detail, giving greater prominence to what has been implemented and highlighting the contribution and innovation with respect others firmware.
Answer:
We have partially reworked the Contributions section, trying to better clarify what we considered to be the important contributions of this paper. In particular we emphasize the importance of the abstract conceptualization our implementation is based upon -- which is also described in more detail in the technical sections (see Figure 2) -- and of the investigation of the tradeoffs between inexpensive hardware and overall scalability.
Reviewer 2 Report
Line 365-371 represents "write type" command?
I suggest, in conclusion, to specify more clearly , possible limits in terms of data volume and number of sensors and actuators in real-time applications, in general.
Author Response
We would like to thank the anonymous reviewers for their comments, which helped us to improve the paper. Following are our answers to specific points raised in this review.
Comment:
I suggest, in conclusion, to specify more clearly , possible limits in terms of data volume and number of sensors and actuators in real-time applications, in general.
Answer:
We reworked the final subsection of the Experiments section, by discussing in more detail limitations and opportunities that emerge from our experimental evaluation.
Reviewer 3 Report
The article presents and discusses a cost effective IoT/IIoT framework architecture for rapid prototyping for industrial and domestic applications.
The application of a cost-effective rapid framework is relevant for the manufacturing industry especially for SME’s. This helps them to embark on Industry 4.0 journey and rapidly adopt advanced technologies -in particular IIoT. Such applications also help manufacturers to see the value of data and how it can be used to optimize the performance of the manufacturing process, enabling the transition into smart factories.
The article is well presented, however, it significantly lacks the related work in the field, especially in IIoT application in industry today and the different/ similar frameworks used.
There are available commercial leading IoT platform that provide an end-to end IIoT solution e.g. AWS IoT & Greengrass, ThingWorks, Cumulocity IoT etc. These types of IoT platforms use similar framework and can communicate with cost-effective embedded systems for data collection and edge processing. Some of these platforms do also offer free-subscription for end users to build end-end IIoT solution for rapid prototyping.
There are also solutions readily available in the market that are based on cost-effective embedded systems, easily configurable and combine multiple built-in open-source lightweight tools for data collection, communication exchange, edge processing and able to connect to external systems and cloud applications. One example is the BB400 of brainboxes which can also be used for rapid prototyping.
The related work section in the article needs to be revisited and expanded further to highlight the current state of IIoT solutions in industry, the frameworks used together with examples of industrial applications of IIoT especially in manufacturing.
The authors need to discuss the novelty of the work further. For example, How does it compare with other frameworks? What’s the advantages of the proposed framework compared to what’s already exists and used in the felid? The authors have highlighted in the abstract the excellent cost to performance ratio based on experimental results of proposed framework. This result also needs to be compared with other framework solutions and discussed further in the discussion section.
Author Response
We would like to thank the anonymous reviewers for their comments, which helped us to improve the paper. Changes to the paper are highlighted in blue. Following are our answers to specific points raised in this review.
Comment:
The article is well presented, however, it significantly lacks the related work in the field, especially in IIoT application in industry today and the different/ similar frameworks used.
There are available commercial leading IoT platform that provide an end-to end IIoT solution e.g. AWS IoT & Greengrass, ThingWorks, Cumulocity IoT etc. These types of IoT platforms use similar framework and can communicate with cost-effective embedded systems for data collection and edge processing. Some of these platforms do also offer free-subscription for end users to build end-end IIoT solution for rapid prototyping.
There are also solutions readily available in the market that are based on cost-effective embedded systems, easily configurable and combine multiple built-in open-source lightweight tools for data collection, communication exchange, edge processing and able to connect to external systems and cloud applications. One example is the BB400 of brainboxes which can also be used for rapid prototyping.
The related work section in the article needs to be revisited and expanded further to highlight the current state of IIoT solutions in industry, the frameworks used together with examples of industrial applications of IIoT especially in manufacturing.
The authors need to discuss the novelty of the work further. For example, How does it compare with other frameworks? What’s the advantages of the proposed framework compared to what’s already exists and used in the felid? The authors have highlighted in the abstract the excellent cost to performance ratio based on experimental results of proposed framework. This result also needs to be compared with other framework solutions and discussed further in the discussion section.
Answer:
We agree that the original Related Work section needed to be improved, and we took this opportunity to completely rework it. We now discuss:
- different hardware platforms for embedded systems, including Espruino and Tasmoda
- different protocols, including XMPP, CoAPP and AMQP
- different architectures, including edge and fog-based.
In addition, we report a comparison to commercial solutions, including:
- data collection and processing platforms, like Thingworx and Cumulocity IoT
- other frameworks closer in spirit to ours, namely Greengrass/AWS IoT and Brainboxes BB400.
We tried to emphasize the main differences and the original contributions made by this paper
Round 2
Reviewer 1 Report
The work has been significantly improved, especially in the introductory part and in comparison with existing solutions.
However, the sentences between lines 189-194 should be supported by citations from relevant works.
Some typos are presents:
- in line 207 there is a typo MQQT instead of MQTT.
- The same MQTT acronymous has must be explained at the first occurrence (line 199) and not at line 205.
Author Response
Thanks for these further comments.
We added two classical references on advantages of pub/sub architectures wrt to client-server, point-to-point ones.
We also fixed typos and anticipated the definition of acronym MQTT.
Reviewer 3 Report
The related work section in the article have been significantly improved together with the discussion and the presentation of the contribution to knowledge. I suggest that the abstract section needs to be improved further by highlighting the main results, the main conclusion of the work and the significant contribution to knowledge.
I also suggest to the authors to include couple of references (if available) from the literature of applied IIoT solutions in industry (that have been included in the compared solutions section or alternative) in order to compare and discuss their performances to the proposed framework. This will make the significant contribution of the article more clear and stronger.
Author Response
Thanks for these further comments.
We have improved the abstract in order to reflect the additional contents added in the review.
Unfortunately, we were not able to find papers describing the adoption of Greengrass or Brainboxes in industrial applications that provided sufficient details on the solution to allow for a comparison. This is not surprising, given the commercial nature of these technologies.
