=Paper=
{{Paper
|id=Vol-2702/sample-10col
|storemode=property
|title=Simplify Node-Red For End UserDevelopment in SeismoCloud
|pdfUrl=https://ceur-ws.org/Vol-2702/EMPATHY_2020_paper_10.pdf
|volume=Vol-2702
|authors=Enrico Bassetti,Edoardo Ottavianelli,Emanuele Panizzi
|dblpUrl=https://dblp.org/rec/conf/avi/BassettiOP20
}}
==Simplify Node-Red For End UserDevelopment in SeismoCloud ==
https://ceur-ws.org/Vol-2702/EMPATHY_2020_paper_10.pdf
Simplify Node-RED For End User Development in
SeismoCloud
Enrico Bassettia , Edoardo Ottavianellia and Emanuele Panizzia
a
Computer Science department, Sapienza University of Rome, Piazzale Aldo Moro 5, 00185 Roma, Italy
Abstract
Networks of IoT devices often require configuration and definition of behavior by the final user. Node-
RED is a flow-based programming platform commonly used for End User Development, but it requires
networking and protocols skills in order to be efficiently used. We add a level of abstraction to Node-RED
nodes in order to allow non-skilled users to configure and control networks of IoT devices and online
services. We applied such abstractions to the SeismoCloud application for earthquake monitoring.
Keywords
End User Development, EUD, Human Computer Interaction, HCI, Internet of Things, IoT, Node-RED
1. Introduction
Smart objects and Internet of Things (IoT) devices are becoming widespread, thus they more
and more often require that end users define their behavior and configure them to connect with
other objects as well as with online services.
Programming tools exist that allow wiring together hardware devices, APIs and online
services, offering a powerful mechanism that provides full control to the end user, but that often
require users to be knowledgeable about technical issues.
Node-RED [1] is a flow-based programming [2] platform, commonly used for End User
Development. It is built as an event-driven platform which routes messages between nodes. A
group of nodes that execute one or more actions starting from an event is called action flow.
These nodes are provided officially by Node-RED team, or they can be added as third-party
plugins. This makes Node-RED highly extendable, while remaining very simple as approach.
However, Node-RED’s “End User” is a person who has some knowledge about underlying
protocols such as MQTT, HTTP requests, WebSockets, etc. In fact, works in literature that use
Node-RED are requiring some prior knowledge of underlying protocols (for example, in Tabaa
et al. [3], skills on Modbus standards and protocols are required).
The purpose of our contribution is to allow non-technical users to use Node-RED to configure
and control networks of IoT devices and online services. This has been shown to be viable using
abstraction with domain specific items[4].
EMPATHY 2020: Empowering People in Dealing with Internet of Things Ecosystems, Sep 29, 2020, Island of Ischia, Italy
email: bassetti@di.uniroma1.it (E. Bassetti); ottavianelli.1756005@studenti.uniroma1.it (E. Ottavianelli);
panizzi@di.uniroma1.it (E. Panizzi)
orcid: 0000-0003-4804-0311 (E. Bassetti); 0000-0002-7442-8451 (E. Panizzi)
© 2020 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
Workshop
Proceedings
http://ceur-ws.org
ISSN 1613-0073
CEUR Workshop Proceedings (CEUR-WS.org)
To reach our goal, we adapted Node-RED to provide the end user with nodes that have a
higher level of abstraction, so that non-technical users can configure and personalize their IoT
ecosystem, still maintaining deep control over it.
2. Related works
Node-RED is widely used as EUD interface for different projects, ranging from simple IoT sensors
integration (like temperature sensors [5] or air quality monitors [6]) to home automation [7], to
complex industrial automation [3]. It is also used to teach in Data Engineering courses [8] due
to its simplicity.
A similar product is IFTTT “if this then that”[9], which is a cloud-based trigger-action End
User Development platform. However, IFTTT supports only simple flows[10] (from triggers to
action), and not complex flows as Node-RED does (for example, it is impossible to synthesize in
IFTTT an industrial automation as in Tabaa et al. [3]).
Ghiani et al.[4] provides an extensive and detailed analysis on the mechanisms behind EUD
when targeting an end-user. Node-RED itself already implemented some of these techniques.
However, while Ghiani et al. “TARE” system is using a composition technique with buttons to
build rules, our proposal uses a visual approach, based on the idea of ”information flow”.
3. Application scenario
The project where we applied Node-RED is SeismoCloud [11], a low-cost earthquake early
warning system, based on smartphones and Internet-of-Things devices. SeismoCloud generates
events from different sources, such as official earthquake feeds, vibrations detected by sensors
deployed by users in a crowd-sensing fashion, and metadata (sensor temperature, status, etc).
It uses these sources internally to store and render data, and to generate earthquake early
warnings.
Node-RED is the core of the new End User Development platform in SeismoCloud, allowing
users to create custom flows1 in order to access and use the data provided by the different
sources above and create personal online services (e.g. alarms, statistics, etc.). As our users do
not generally have computer science knowledge, our goal is to provide a tool that everyone can
use, without requiring technical skills. However, to reach this level of abstraction, we had to
modify Node-RED’s “tool set”, removing any needs of knowledgeability and understanding of
IoT- and network-related concepts.
4. Adding abstractions in Node-RED
A SeismoCloud sensor, just like the majority of IoT sensors, uses MQTT for signaling/data. Even
if MQTT is a very simple protocol, all specific configuration about the MQTT connection, topic
subscription and publishing may be out of range for a normal user. Also, some information
1
Node-RED user can drag and drop nodes into one or more flows (group of nodes for specific purposes, e.g.
room light control), and link the output of a node to the input of another node allowing a message “produced” by a
node to be “consumed” by the linked one (which, in turn, can output another message).
(a) Temperature output (b) Perceptible earth-
node quakes (per sensor)
Figure 1: Example of two new nodes created for this experiment
Figure 2: A very simple flow in Node-RED. Messages flows from left to right.
about earthquakes are available via REST API calls (for mobile apps and website, mostly). These
technologies require at least some programming skills to be used.
In order to solve this problem, we created some domain-specific nodes that users can drag
and drop and configure just specifying domain-specific information in human-readable form.
These nodes are abstractions for the Node-RED supported technologies, like MQTT or REST. For
example, the “temperature” node (fig. 1a) can be easily configured just specifying the IoT sensor
name, without any knowledge of the MQTT configuration which is automatically performed.
In fact, when the IoT sensor sends its temperature using MQTT to all its subscribers, this
value is received by all the “temperature” nodes (fig. 1a) configured using its name. All details
such as MQTT broker, credentials, topic name, QoS, TLS configuration, etc., are hidden from the
point of view of the user.
This simplifies the interaction with Node-RED and allows anyone to use its EUD interface.
By mixing Node-RED standard nodes and our high level abstraction non-skilled people can
build and understand Node-RED flows, such as “when two devices emits a vibration message,
send an SMS” (see figure 2).
5. User testing and final results
To evaluate whether our approach represents a sufficient simplification for standard users and
an enabler for the EUD in SeismoCloud, while maintaining enough expressive power to create
flows as desired by users, we planned and executed a test with 7 non-skilled users plus 1 expert
user.
Users were asked to execute 4 tasks with different level of difficulty which required to get
a data source event (e.g. a seismometer vibration detection or a temperature sensor above-
threshold) and trigger a message using a simple label-based template. They were expected to
drag and drop the proper nodes from the palette to the workspace and create a Node-RED flow.
For all trials we recorded the duration, success or failure in achieving task goals, whether the
user understood newly introduced nodes. Also, users were asked to think aloud.
After performing tests with the first 4 users we made a second version of the nodes, simplifying
the labels that describe the information required for their configuration. This problem was
found asking the users if they were understanding what they were doing. All the non-skilled
users replied they understood that node returned a value, but the meaning wasn’t clear.
In the following tests, all the users were able to complete the tasks without significant
problems, although sometimes it took more time than we expected.
The result of this test is encouraging. We are currently planning a test with expert users from
our network of beta testers, in order to asses if the current interface is expressive enough and
to what extent the underlying Node-RED built-in nodes are still necessary for operation in the
SeismoCloud environment.
References
[1] I. E. Technology, Node-red, 2020. URL: https://nodered.org.
[2] J. P. Morrison, Flow-Based Programming: A new approach to application development,
CreateSpace, South Carolina, US, 2010.
[3] M. Tabaa, B. Chouri, S. Saadaoui, K. Alami, Industrial communication based on modbus
and node-red, Procedia computer science 130 (2018) 583–588.
[4] G. Ghiani, M. Manca, F. Paternò, C. Santoro, Personalization of context-dependent appli-
cations through trigger-action rules, ACM Transactions on Computer-Human Interaction
(TOCHI) 24 (2017) 1–33.
[5] M. Lekić, G. Gardašević, Iot sensor integration to node-red platform, in: 2018 17th
International Symposium INFOTEH-JAHORINA (INFOTEH), IEEE, IEEE, Piscataway, NJ,
2018, pp. 1–5.
[6] S. Chanthakit, C. Rattanapoka, Mqtt based air quality monitoring system using node mcu
and node-red, in: 2018 Seventh ICT International Student Project Conference (ICT-ISPC),
IEEE, IEEE, Piscataway, NJ, 2018, pp. 1–5.
[7] A. Rajalakshmi, H. Shahnasser, Internet of things using node-red and alexa, in: 2017
17th International Symposium on Communications and Information Technologies (ISCIT),
IEEE, IEEE, Piscataway, NJ, 2017, pp. 1–4.
[8] Z. Chaczko, R. Braun, Learning data engineering: Creating iot apps using the node-red and
the rpi technologies, in: 2017 16th International Conference on Information Technology
Based Higher Education and Training (ITHET), IEEE, IEEE, Piscataway, NJ, 2017, pp. 1–8.
[9] S. Ovadia, Automate the internet with “if this then that”(ifttt), Behavioral & social sciences
librarian 33 (2014) 208–211.
[10] B. Ur, E. McManus, M. Pak Yong Ho, M. L. Littman, Practical trigger-action programming
in the smart home, in: Proceedings of the SIGCHI Conference on Human Factors in
Computing Systems, 2014, pp. 803–812.
[11] E. Panizzi, The seismocloud app: Your smartphone as a seismometer, in: Proceedings of the
International Working Conference on Advanced Visual Interfaces, ACM, 1601 Broadway,
Times Square, New York City, 2016, pp. 336–337.