Executing IoT Processes in BPMN 2.

0:
Current Support and Remaining Challenges
Faruk Hasić and Estefanı́a Serral Asensio
Department of Information Management, Modelling and Simulation
KU Leuven, Warmoesberg 26, 1000 Brussels, Belgium
[email protected]; [email protected]

Abstract—The Internet of Things (IoT) has seen a vast surge and academia. Furthermore, BPMN 2.0 adheres to the model-
in recent years, both in academia and in industry. Business driven development (MDD) paradigm, which recognises that
processes are starting to take advantage of IoT to interact with models rather than computer programs are the primary focus
the increasingly connected and dynamic physical environment.
Recent literature has focused on analysing how this interaction and drive of software development. As such, models need to
between processes and IoT devices can be modelled, i.e. how be detailed enough to be executable. Though detailed, models
the process can be described for discussion purposes. However, should be abstract of the technology used, as the technology
little attention is given to configuring the execution of these should be made explicit during the execution of the system.
IoT-based processes. IoT-enhanced processes should indeed be
This paper is structured as follows. Section II provides
able to incorporate IoT devices in the endeavour of process
execution. This paper discusses research challenges in IoT process background information on process execution, the Internet of
configuration and execution using the executable Business Process Things, the link between IoT and business processes, and the
Model and Notation 2.0 (BPMN 2.0) standard. As such, we related work on the execution of IoT processes. Section III
constitute an agenda for future research. provides a real-life IoT healthcare running example which
Keywords. Internet-of-Things, IoT, Internet-of-Things Processes, will be used to illustrate the challenges. Section IV discusses
IoT Processes, Process Execution, IoT Process Execution, Process which elements of BPMN 2.0 are suitable for IoT process
Automation, IoT Process Automation.
execution. In Section V we identify a set of challenges that
I. I NTRODUCTION need to be tackled to guarantee the correct execution of
The network of interconnected devices that constitute the IoT-enhanced processes, and we identify works that partially
IoT keeps growing. A plethora of devices are used to facilitate address these challenges. Finally, Section VI concludes and
both businesses and individuals in their daily endeavours. provides directions for future research.
The application areas are both vast and diverse, ranging from
smart home environments [1] to medical monitoring systems II. BACKGROUND AND R ELATED W ORK
[2]. Usually, a single device does not suffice to encapsulate In this section, we give an overview of concepts and
the physical complexity and therefore multiple devices are research areas that are relevant for the execution of IoT-
combined to render a system that accommodates the reality. enabled business processes.
Business processes can benefit from IoT by interacting with
these devices to steer their behaviour. Ideally, the business pro- A. The IoT paradigm
cess is in turn capable of adapting to its execution environment
by tailoring the process to case-specific needs as demanded IoT refers to the ever-growing network of interconnected
by the context. This integration of IoT and business processes devices. Generally speaking, an IoT environment is populated
has enjoyed significant attention in literature, and in particular by two kinds of devices: sensors and actuators. Sensors serve
when it comes to the integrated modelling of processes and as a mechanism to capture reality and are therefore the
IoT devices [3]. This interaction between IoT devices and source of IoT data towards the system. Actuators on the other
the business process remains of paramount importance during hand operate in the reverse direction, i.e. they change the
the execution of the process as well. However, very few reality by activating a physical reaction that is initiated by
researchers have investigated this topic, while only providing the system. Typical sensors are for example humidity sensors,
partial solutions. In this paper, we approach that gap by light sensors, camera sensors, or heart rhythm sensors. On the
contributing an incomplete yet indispensable set of research other hand, typical actuators can be an electric motor, a heating
challenges that need to be tackled to achieve IoT process actuator, a sprinkler system, or a speaker. Data collected by
execution using the Business Process Model and Notation 2.0 IoT devices is considered to be part of the context information
(BPMN 2.0) standard [4]. We focus on BPMN 2.0 since it [5], [6], i.e. the information describing the conditions under
is an executable and widely accepted standard in industry which the system operates. Relevant context information might
for instance include the location of an object, the humidity
level, and the temperature. Thus, IoT devices aid in digitising
978-1-7281-4844-1/19/$31.00 2019
c IEEE the context that manifests itself in the physical environment.
B. Business processes and their execution exist that can collect patient data relevant for the patient
The field of Business Process Management (BPM) is con- monitoring process [2]:
cerned with defining and executing business activities within – Electrocardiogram (ECG) sensors monitor the heart.
an organisation. The more context-intensive the environment – Respiratory sensors check the breathing rate.
in which the processes takes place, the more challenges arise – Skin temperature sensors monitor the skin temperature.
regarding the modelling and execution of said processes. A – Muscular Electromyography (EMG) sensors monitor the
process can be specified and executed using the BPMN 2.0 muscle activity.
standard. The models expressed in BPMN 2.0 and saved in – Blood pressure sensors measure the blood pressure.
XMI [7], are interpreted by a process engine, like Camunda All these sensors collect measurements on the patient’s
[8], to execute the corresponding tasks as specified in the health. IoT settings are often data- and decision-intensive. Note
models. that a single sensor or even a few sensors combined are not
enough to capture the COPD. For instance, the patient might
C. IoT business processes
take a walk outside in the winter and a sensor registers a low
Recent literature has focused on incorporating IoT into skin temperature. In that case, the patient is not necessarily
BPM [3], [9]. IoT and business process integration can be suffering from COPD at that moment. However, an expert can
twofold. By definition, IoT business processes are enacted build patient-specific decision rules to capture COPD in such
in a dynamic and highly connected physical environment. a monitoring system. For instance, if the sensors register a low
Since IoT is a technology that can be used to digitalise the skin temperature, a short and fast breathing rate, together with
context of a system, the process is granted the capability of high blood pressure, the monitoring process might decide that
both understanding its context and of changing it through IoT the patient is suffering an attack and running out of oxygen. In
devices. This was also firmly stressed in [3], indicating that such a situation the process can trigger the administration of an
IoT technology can help concretising abstract process models oxygen mask to the patient. An illustrative COPD monitoring
and help the processes deal with new situations. Therefore, on IoT process is given in Figure 1. First the heart rhythm of
one side IoT devices can be used to digitalise the context of a the patient is checked and the emergency alarm is activated if
business process and consequently to incorporate the context needed. Afterwards, the severeness of the COPD is assessed,
into the process execution. That way, an abstract process model and given the assessment, the process ends or treatments with
can be made concrete for context-aware execution [3]. On oxygen masks or inhalers are carried out.
another side, IoT devices can be used to automate tasks. This
can be a physical task, such as opening the curtains in a IV. BPMN 2.0 S UPPORT FOR I OT P ROCESS E XECUTION
smart home, or a digital task, such as sending a message or To achieve IoT process execution, IoT devices must be
notification to a system or a human user. linked to process elements as described in Section II-C. The
Most papers on the cross section of IoT and process BPMN 2.0 standard offers different mechanisms that could be
management mainly deal with the integrated modelling of used to establish this link. We can group these mechanisms as
IoT devices in business process models. Examples can be follows:
found in [10]. Few works, however, do touch on the subject
1) Tasks. The latest BPMN standard defines three types of
of configuring IoT processes for execution, e.g. [1]. Most of
automated tasks:
them rely on code generation, such as [11]. Others use existing
process engines for model interpretation, while resorting to a) Script tasks have scripts attached to them that are
code generation for the concepts not supported by the process written in a language that the process engine can parse
engine [12]. Another approach is to avoid code generation and consequently execute in an automatic fashion.
and to extend an existing process engine to support the newly b) Service tasks are tasks that invoke an application or
introduced constructs [13], or to develop an entirely new service to automatically complete the task. Since IoT
custom process engine [14]. devices can be managed by IoT services that are
invocable by other systems, the service tasks are a
III. RUNNING E XAMPLE OF AN I OT P ROCESS suitable mechanism to link the IoT infrastructure to
Consider a patient health monitoring system for a person the executable business process. The service task can
diagnosed with the Chronic Obstructive Pulmonary Disease forward a method call to, for instance, a Java delegate
(COPD). COPD is a disease that obstructs the lungs and ob- class, which in turn implements an IoT service. That
structs the airflow and breathing of the patient. Acute attacks of way, communication and data propagation between IoT
the disease can happen. In that case the patient can experience devices and the business process can be implemented.
uncomfortable complications such as fast breathing, a fast c) Business rule tasks can be used to incorporate IoT
heart rate, hyperactive use of muscles, and a cold skin [2]. resources into the business process. IoT data collected
In scientific literature it has been recognised as well that an by the IoT devices can have the role of input data to
IoT-based patient monitoring process can help increase the life the decision logic hidden behind a business rule task.
quality of the patient and decrease the risks that are inherent The business rule task can possibly invoke a separate
to the disease [2]. Multiple sensors and wearable technologies decision model that encapsulates the logic [15], [16].

2
 Sound
Administer
emergency
oxygen mask
alarm Severe
attack

Check heart OK Check COPD OK

Mild
attack Administer
inhaler

Figure 1: A process model for a COPD monitoring IoT process.

Figure 2 depicts the task types that are suitable for mod- to execute the IoT process.
elling IoT processes, i.e. a script, service, and business 4) Data. Notice that the data interchange between the IoT
rule task according to the BPMN 2.0 standard [4]. devices on the one hand and the IoT process on the other
2) Events. Another way to establish communication be- is of paramount importance in a highly interconnected
tween IoT devices and the business process is the use and dynamic setting where the IoT data drives the deci-
of events that are triggered by an IoT device. If an IoT sions made in the process. Such a setting was illustrated
device registers an action in the physical world, this in the running example in Section III, where the decisions
is reflected in the process by triggering an event. The in the process are driven by a data collected from multiple
execution of the event can subsequently trigger a part IoT devices. The IoT infrastructure produces plenty of
of the process. Similarly, the process can trigger an IoT data that can be relevant for the process. This data
device if a certain event manifests itself. needs to be interpreted and correctly utilised by the IoT
3) Resources. Note that the IoT devices can be viewed as process. Thus, IoT processes are strongly data-intensive
resources of an IoT process. According to the BPMN 2.0 as the data perspective of IoT processes revolves around
standard specification [4], the resource class in the meta data elements being consumed, produced and exchanged
model is used to specify resources that can be referenced during process execution. This emphasises that a critical
by activities. These Resources can be human resources, resource for an IoT process is the IoT data itself.
but also any other resource assigned to activities during
process execution. Furthermore, the standard defines re- V. C HALLENGES
sources as abstract, i.e. it only defines the resources with- Although the BPMN 2.0 standard already offers good
out specifying which specific resource is allocated to the support for different aspects of IoT process execution, some
process element. Actual resource allocation is resolved challenges still remain. Here, we enumerate a set of research
at runtime. Given that IoT devices can be viewed as challenges that need to be addressed to guarantee a sound IoT
resources, the integration of IoT technology and business process execution, relying on the BPMN 2.0 standard. For each
processes can be approached in a BPMN 2.0 compliant challenge we follow the same structure: we briefly describe
manner. Consider IoT devices as resources which can the challenge, provide an illustration based on the running
be invoked from the process model. Since the standard example in Section III, and describe relevant literature that
specification considers abstract resources, this can be tries to approach the research gap described in the challenge.
achieved without administering changes or extensions An overview of the challenges is provided in Table I.
to the standard BPMN specification. Furthermore, the
A. Resource allocation challenges
existing BPMN 2.0 elements with automation properties,
such as service tasks and business rule tasks, can be used Inherent in IoT processes is the use of IoT devices as
resources for task automation and context adaptation. In what
follows, we point out challenges for IoT resource awareness
in processes.
Challenge 1 (C1): IoT resource-awareness in tasks
Script Service Business rule
task task task 1) Challenge description: In the previous section, we dis-
cussed the current support of the BPMN language towards
incorporating IoT devices in the business process from a
Figure 2: BPMN 2.0 script, service, and business rule tasks. modelling perspective. The challenge is to configure the IoT

3
process in a technology-independent way, yet with a sufficient Challenge 2 (C2): IoT resource binding at runtime
level of detail to interact with the available resources at 1) Challenge description: In an IoT infrastructure, multiple
execution time. devices that monitor or execute the same physical reality can
2) Example: Take for instance the sound emergency alarm be put in place. In such a setting, rather than a straightforward
service task of the running example in Figure 1. The service one-to-one mapping of IoT resources and process elements,
task is linked to the alarm device in the connected reality. The as described in the previous challenge, a selection of the
execution of the alarm device corresponds to the execution of proper resource at runtime is necessary. This link between
the sound emergency alarm service task. Here it is clear that the process model and the IoT devices should be established
the business process is guiding the use of the IoT technology depending on certain criteria, such as device availability,
in the physical world. However, as previously mentioned, the accuracy, malfunctioning, and so on.
other direction of integrating IoT devices and the business 2) Example: Take for instance the sound emergency alarm
process is a possibility as well. Namely, the IoT devices can service task of the running example in Figure 1 which is linked
be used to digitalise the context in which the business process to an alarm device in the connected reality. The most relevant
is executed. This can for instance happen through the check alarm device can be selected for execution, e.g. the speakers
heart rhythm and check COPD severeness business rule tasks present on the floor of the hospital where the patient is, or the
in the process model in Figure 1. The logic hidden behind the smart phone of the physician that monitors the patient if the
business rule tasks is dependent on data collected from the physician is currently not on the same floor.
IoT resources. 3) Possible solutions: Optimised resource allocation is con-
3) Possible solutions: Researchers have argued that BPMN sidered an important challenge in the merger of IoT technolo-
2.0 is capable of modelling IoT processes without necessarily gies and business processes [3]. Among other solutions, the
extending the existing standard specification. For instance, [11] logic on which device to select at runtime can be modelled
shows how BPMN elements can be used to model Wireless in a decision model. A standard suitable for such decision
Sensor Network Applications. The devices and actions per- modelling is the Decision Model and Notation (DMN) [18]
formed with the devices are not represented in the model. provided by the Object Management Group. This model can
Thus, no discrimination between regular automated tasks and be invoked at runtime to dynamically bind an IoT device to
IoT tasks is present in these models, both when it comes to the relevant process element. That way, a more efficient IoT
process modelling as well as process execution. The service resource utilisation can be achieved and process execution can
tasks simply call upon a service responsible for managing be safeguarded. A similar approach that aims at resolving IoT
an IoT device. Others, however, have indicated the need to resource conflicts in business processes is suggested in [19].
distinguish between regular automated tasks and IoT tasks Here, the authors define event condition action (ECA) state-
in the design phase. They propose a visual discrimination ments indicating when and under which conditions an already
between regular automated tasks and IoT tasks, as well as allocated resource should be substituted with an equivalent
between regular events and IoT events [9]. They link IoT resource.
resources to the process by adding a new symbol indicating Challenge 3 (C3): IoT resource malfunction event handling
that an activity or event is linked to an IoT resource, and they 1) Challenge description: Once IoT resources have been
also extend the meta model to include the newly proposed allocated to a process element, be it by a simple one-to-
types of IoT tasks and events. These newly defined process one mapping as discussed in C1 or by dynamic resource
elements can be made executable through code generation, binding at runtime as considered in C2, the IoT resource may
by extending existing process engines to incorporate the new still be deficient or inaccessible during the enactment of the
elements, or by developing a custom process engine. Note relevant process elements. These unexpected events should be
that such an approach deviates from the standard BPMN 2.0 captured in the process. In case an allocated IoT device is
support as described in Section IV. Hence, a trade off exists not responding or failing, the IoT process might not be able
between conforming to the standard specification or extending to continue its execution. Such failures should be noticed and
it for IoT purposes. The former guarantees that the IoT process captured in the process. Resource failure monitoring is crucial
remains executable according to BPMN 2.0, while the latter for the quality of task execution [3].
explicitly models IoT devices in the process. 2) Example: In the running example of Figure 1, such a
Approaches modelling the IoT device within the business safeguard should be foreseen at service task sound emergency
process are straightforward as they resort to a one-to-one alarm. If the alarm device fails during execution, the conse-
mapping between process elements and IoT devices. However, quences for the patient being monitored are potentially severe.
a one-to-one mapping between an IoT device and a process 3) Possible solutions: In the case described in the example,
element is not necessarily satisfactory. Furthermore, IoT pro- opting for process termination is not an option. Therefore,
cesses can rely on multiple IoT devices, even within the same a routine should be put in place to select a different alarm
process element. Therefore, rule-based IoT resource allocation device or to convey the message to the physician in some other
should be investigated, in analogy with works on regular manner, e.g. by sending a text message or smart notification to
resource allocation and scheduling in business processes [17]. the physician’s smart phone or by initiating an automated call.

4
In case of multiple alarm devices, by re-invoking the service can for instance be done by using a service task which queries
task responsible for the operation of the failing device, an the database and retrieves the necessary data. This data can
alternative device can be dynamically bound to the service subsequently be used in the following stages of the process.
task. A modelling pattern should be foreseen in the process Additionally, a combination of push and pull is possible as
model to capture failures and to call upon another IoT device well, i.e. some IoT devices store data in the database such
to perform the execution. Remedying resource malfunction that the process can pull data on demand. Other IoT devices
complications can be done by for instance setting up error immediately notify the process by for instance triggering an
event handling modelling patterns to assure that processes event in the process that captures the sent data and provides
can continue, terminate, restart, or that they can be partially it to the process for further interpretation.
rolled back to reiterate the failed activities with other available Existing work on the integration of ontologies and task mod-
resources. In [11], some error handling patterns for wireless els [21], [22] in IoT environments, as well as the integration of
sensor networks (WSN) have been introduced. The patterns ontologies and Petri Nets [1] can provide inspiration towards
rely on events that capture undesirable behaviour. Another solving the data retrieval challenge. Furthermore, relevant
interesting line of research are self-healing business processes. works also include those that provide possibilities to retrieve
Here, the idea is to replace at runtime services that fail during data based on semantic queries over ontologies [23].
process execution. An example is provided in [20], where B. Communication challenges
the authors define monitoring and recovery actvities to ensure
process resilience. Introducing IoT resources to the business process invokes
additional questions regarding the communication between
Challenge 4 (C4): IoT data retrieval: push vs. pull the business process and said resources. We elaborate further
1) Challenge description: Another important challenge to below.
consider is understanding the context situation in which the Challenge 5 (C5): IoT and human resource communication
process is being executed. This can be done by pushing the 1) Challenge description: Regular non-automated processes
data to the process from the IoT devices or by the process tasks are usually performed by human actors. Whenever
pulling the data from the devices on demand. In the former users finish a task that has been assigned to them, they
case a sensor performs a data push to the process. However, indicate this by sending a confirmation to the system. This
given that IoT devices often provide a stream of data, the data way, the next task in line becomes enabled and the human
pushes are likely to happen frequently. resource responsible for the enactment of that task can start
2) Example: In the running example of Figure 1, the data with the execution. Notice, however, that in an IoT setting
necessary for invoking the business rule activities check heart task automation is a commonplace pattern. Hence, certain
rhythm and check COPD severeness can be pulled from the tasks will not be executed by human resources, but rather
IoT sensors described in the running example in Section III. by the system itself. This contributes to the co-existence
Notice however, that a data push from the IoT devices to of different types of resources in an IoT process [3]. In
the process might be required as well in certain situations. such a setting, the system needs to notify the process that
If the heart rhythm sensor notices a steep increase of the heart the automated task has been executed successfully such that
rhythm which is followed by a quick decrease in a short period subsequent tasks, be they human tasks or automated system
of time, this might point towards heart arrhythmia. If the heart tasks, become enabled. If the subsequent tasks are driven by
rhythm sensor captures such behaviour, the IoT device should human resources, the system should notify the users about
push a notification to the process. A data push from the IoT system task completion and it should indicate that the tasks
device to the process can for instance be captured in the the human resources are responsible for are permitted to be
process through an event, which in turn can trigger activities executed. Thus, a message notification system should be put in
that update the patient’s medical file and that automatically place to facilitate the interaction between automated IoT tasks
warn a responsible physician. and human user-dependent tasks. Furthermore, these message
3) Possible solutions: As the example illustrates, a trade notifications should be organised in such a way that they
off between pushing IoT data to the process and providing the maximise effectiveness, i.e. that they do not needlessly bother
process with IoT data on demand should be made when setting the receivers of the notifications and that they do not produce
up the system to guarantee effective process enactment. The non-value adding overhead tasks for the human users.
data can be captured in the process by, for instance, events that 2) Example: In the running example of Figure 1, the sound
are triggered when the IoT device pushes data to the process. emergency alarm activity might be executed for a patient and
Another possibility is for the process to pull the data captured a nurse responsible for the manual activity administer oxygen
by the devices on demand. The devices register data and store mask starts a shift after the alarm, or the nurse was not able
it into a database or knowledge base, if the most current data to hear the alarm at the time. In that case the nurse should be
update by the devices is considered relevant. In such a way, notified to indeed administer the oxygen mask to the patient,
the data stored in the database or knowledge base is the most but to be careful not to move the patient into a bad position
recent relevant data. The process can then pull necessary data given the abnormal heart rhythm as established in the check
from the database whenever it is needed in the process. This heart rhythm decision activity.

5
 1) Possible solutions: In [24] simple message flows in [4] OMG, “Business process model and notation (BPMN) 2.0,” 2011.
BPMN between IoT tasks and human tasks have been pro- [5] G. Hermosillo, L. Seinturier, and L. Duchien, “Creating context-adaptive
business processes,” in International Conference on Service-Oriented
posed to facilitate co-existent resource communication. A Computing. Springer, 2010, pp. 228–242.
more advanced approach can be found in [25], where the [6] J. Hu, G. Aghakhani, F. Hasić, and E. Serral, “An evaluation framework
authors propose considerate notifications which interact with for design-time context-adaptation of process modelling languages,” in
IFIP Working Conference on The Practice of Enterprise Modeling.
users in a way that is not disturbing for the user. They provide Springer, 2017, pp. 112–125.
a possibility to personalise and adjust the level of obtrusiveness [7] OMG, “Xml metadata interchange (XMI) 2.5.1,” 2015.
of the notifications. [8] Camunda, “Process engine,” https://docs.camunda.org/manual/7.8/user-
guide/process-engine/, 2018, accessed: 2018-11-16.
Research challenges in BPMN 2.0 IoT process execution [9] S. Meyer, A. Ruppen, and C. Magerkurth, “Internet of things-aware pro-
Resource allocation challenges cess modeling: integrating iot devices as business process resources,” in
C1: IoT resource-awareness in tasks International conference on advanced information systems engineering.
C2: IoT resource binding at runtime Springer, 2013, pp. 84–98.
C3: IoT resource malfunction event handling [10] N. Brouns, S. Tata, H. Ludwig, E. S. Asensio, and P. Grefen, “Modeling
C4: IoT data retrieval: push vs. pull iot-aware business processes-a state of the art report,” arXiv preprint
Communication challenges arXiv:1811.00652, 2018.
C5: IoT tasks and human resources communication [11] A. Caracaş and T. Kramp, “On the expressiveness of bpmn for modeling
wireless sensor networks applications,” in Business Process Model and
Table I: Research challenges in IoT process execution. Notation, R. Dijkman, J. Hofstetter, and J. Koehler, Eds. Berlin,
Heidelberg: Springer Berlin Heidelberg, 2011, pp. 16–30.
VI. C ONCLUSION AND F UTURE W ORK [12] D. Domingos and F. Martins, “Using bpmn to model internet of things
behavior within business process,” IJISPM-INTERNATIONAL JOUR-
This paper discusses the configuration and execution of NAL OF INFORMATION SYSTEMS AND PROJECT MANAGEMENT,
IoT-enhanced processes. More specifically, we investigated to vol. 5, no. 4, pp. 39–51, 2017.
[13] S. Appel, P. Kleber, S. Frischbier, T. Freudenreich, and A. Buchmann,
which extent the latest BPMN standard supports IoT process “Modeling and execution of event stream processing in business pro-
execution. To manage this, we consulted existing works on cesses,” Information Systems, vol. 46, pp. 140–156, 2014.
the integration of IoT and processes and we concluded that [14] T. Peng, C. Chi, A. Chiasera, G. Armellin, M. Ronchetti, C. Matteotti,
C. Parra, A. O. Kashytsa, and A. Varalta, “Business process assignment
most research revolves around the modelling of IoT devices in and execution in mobile environments,” in 2014 International Confer-
processes. Only a select group of researchers also touch on the ence on Collaboration Technologies and Systems (CTS), May 2014, pp.
subject of configuring and executing the IoT-enhanced process 267–274.
[15] F. Hasić, J. De Smedt, and J. Vanthienen, “A service-oriented archi-
models by either relying on already existing process engines, tecture design of decision-aware information systems: Decision as a
custom process engines, or by utilising code generation. As service,” in OTM Confederated International Conferences ”On the Move
most researchers rely on BPMN 2.0 for the modelling effort, to Meaningful Internet Systems”. Springer, 2017, pp. 353–361.
[16] F. Hasić, J. De Smedt, and J. Vanthienen, “Augmenting processes with
we argue that sticking to the core of the standard is beneficial decision intelligence: Principles for integrated modelling,” Decision
for IoT process execution as well. However, we identified a Support Systems, vol. 107, pp. 1 – 12, 2018.
set of challenges regarding IoT process configuration and exe- [17] P. Wohed, W. M. van der Aalst, M. Dumas, A. H. ter Hofstede, and
N. Russell, “On the suitability of bpmn for business process modelling,”
cution according to BPMN 2.0. Furthermore, we analysed the in International conference on business process management. Springer,
literature to identify approaches that address these challenges. 2006, pp. 161–176.
However, the approaches presented in literature are either [18] OMG, “Decision Model and Notation (DMN) 1.2,” 2018.
[19] K. Suri, W. Gaaloul, A. Cuccuru, and S. Gerard, “Semantic framework
not standardised or not sufficient to tackle the challenges. for internet of things-aware business process development,” in Enabling
Therefore, we appeal that these challenges need to be fully Technologies: Infrastructure for Collaborative Enterprises (WETICE),
remedied to obtain a clear view on IoT process execution. 2017 IEEE 26th International Conference on. IEEE, 2017, pp. 214–
219.
The identified challenges should be addressed in the future [20] L. Baresi, S. Guinea, and L. Pasquale, “Self-healing bpel processes
with the aim to solidify the integration of IoT and business with dynamo and the jboss rule engine,” in International workshop
processes for execution. Particular attention will be given on Engineering of software services for pervasive environments: in
conjunction with the 6th ESEC/FSE joint meeting. ACM, 2007, pp.
to alleviate the burdens specified in the research challenges 11–20.
in this paper. More specifically, we will focus on dynamic [21] E. Serral, P. Valderas, and V. Pelechano, “Context-adaptive coordination
IoT resource binding at runtime. Additionally, we will study of pervasive services by interpreting models during runtime,” The
Computer Journal, vol. 56, no. 1, pp. 87–114, 2012.
change patterns for IoT process evolution at runtime. [22] ——, “Addressing the evolution of automated user behaviour patterns
by runtime model interpretation,” Software & Systems Modeling, vol. 14,
R EFERENCES no. 4, pp. 1387–1420, 2015.
[1] E. Serral, J. De Smedt, M. Snoeck, and J. Vanthienen, “Context-adaptive [23] D. Calvanese, B. Cogrel, S. Komla-Ebri, R. Kontchakov, D. Lanti,
petri nets: Supporting adaptation for the execution context,” Expert M. Rezk, M. Rodriguez-Muro, and G. Xiao, “Ontop: Answering sparql
Systems with Applications, vol. 42, no. 23, pp. 9307–9317, 2015. queries over relational databases,” Semantic Web, vol. 8, no. 3, pp. 471–
[2] M. Hassanalieragh, A. Page, T. Soyata, G. Sharma, M. Aktas, G. Mateos, 487, 2017.
B. Kantarci, and S. Andreescu, “Health monitoring and management [24] R. Pryss, M. Reichert, A. Bachmeier, and J. Albach, “Bpm
using internet-of-things (iot) sensing with cloud-based processing: Op- to go: Supporting business processes in a mobile and sensing
portunities and challenges,” in 2015 IEEE international conference on world,” in BPM Everywhere, 2015, pp. 167–182. [Online]. Available:
services computing (SCC). IEEE, 2015, pp. 285–292. http://dbis.eprints.uni-ulm.de/1153/
[3] C. Janiesch, A. Koschmider, M. Mecella, B. Weber, A. Burattin, [25] E. Serral, M. Gil, P. Valderas, and V. Pelechano, “Automating unobtru-
C. Di Ciccio, A. Gal, U. Kannengiesser, F. Mannhardt, J. Mendling et al., sive personalized services in ambient media environments,” Multimedia
“The internet-of-things meets business process management: mutual tools and applications, vol. 71, no. 1, pp. 159–178, 2014.
benefits and challenges,” arXiv preprint arXiv:1709.03628, 2017.