The deployments of IoT technologies would considerably effect
and modify the way in which businesses do business and relations amid diverse segments
of the society, having an effect on several processes. For being capable to gain
the numerous latent profits that are proposed for the IoT, numerous challenges concerning
the modelling as well as implementation of these kinds of processes is required
to be resolved to comprehend broader and
especially profitable deployments of IoT.
There is a need to take into account the exceptional features of IoT
services as well as processes, plus it is probable that current business
process modelling as well as execution languages and service description
languages like USDL, will be required to be protracted.
3.2.2. Adaptive and
It should be noted that the processes get more accustomed to
what is really taking place in the actual world, which is one of the chief
benefits of IoT integration. Characteristically, it is founded on happenings which
are identified either directly or through real-time study of sensor data. These
kinds of events can take place in the process at any time. For certain events,
the probability of occurrence is extremely low, which means that it is known
that they may take place, however not at when time or even they will take place.
Modelling these kinds of events in a process is burdensome, because
these events would be needed to be encompassed in all probable activities, resulting
in added difficulty as well as making it additional hard to comprehend the
modelled process, specifically the chief process flow. Furthermore, the context
will be taken into account to decide how to respond to a lone event, viz. the
events which have been perceived formerly.
Study on event-driven as well as adaptive processes in IoT
systems can reflect the extension as well as use of EDA () for activity checking
as well as complex event processing (CEP). To start particular parts or steps
of a business process, Event Driven Architectures can be joined with business
process execution languages.
Dealing with Unreliable Data
While working with events that are approaching from the
physical world (like from signal processing algorithms or sensors), some amount
of undependability as well as doubt is put in the processes. In case, the events
which have certain doubt in them are used as a basis to take the conclusions in
a business process, it is sensible to link all these events with certain value
for quality of information (QoI).
In easy cases, this lets the process modeller to describe
thresholds – for instance, the event is presumed to be actually taken place if
the degree of certainty is above 90%. In case, it is amid 50% – 90%, certain
other activities would be started to find out if the event took place or not. The
event is disregarded, in case, it is under 50%. In cases of involvement of
numerous events, things get more difficult – for instance, degree of certainty of
one event 52%, one has 95%, and another has 73%. The principal services which
fire the actual events need to be programmed to link these kinds of QoI values
with the events.
From the viewpoint of a BPM, it is vital that this kind of
information is able to be taken, treated as well as articulated in a language
like BPMN, which is a modelling notation language. Furthermore, there is a need
for semantics and syntax of these kinds of QoI values to be standardized. As in
the examples given above, is it merely a certainty percentage, or it must be
more expressive (like a range in which true value is there)? Applicable methods
must not merely look in the uncertainty flow of a business process based on IoT,
however in the total modelling as well as structuring of (perhaps unstructured
or unknown) process flows as well. Practices for uncertain modelling of
processes and data can be taken into account.
Dealing with Unreliable Resources
The data from resources is characteristically undependable,
moreover, the resources offering the data are also undependable, for instance, owing
to the hosting device’s failure. Processes depending on these kinds of
resources require to be capable of adapting to these kinds of conditions. The major
subject is to identify this kind of a failure. The detection in which a process
is calling a resource openly, is unimportant. It is tougher in the case we are
talking regarding resources which may produce an event at one point in time
(for instance, the resource which observers the temperature in the truck, plus gives
an alert, in case, it has got extremely hot). Resource failure can be a reason
for not getting any event, however one other reason can be that nothing was there
to report. Similarly, the generated reports’ quality must be frequently examined
for accuracy. To sense these kinds of problems, certain monitoring software is required;
however, it is not clear if this kind of software must be a separate component
or must be a segment of the BPM execution environment. Amid the challenges of
the research is the management of processes of monitoring with run-time
actuating processes, provided that management planes have a tendency to function
at diverse time scales from IoT processes.