SEPL: An IoT Platform for Value-added Services in the Energy
Domain
Architectural Concept and Software Prototype
Theo Zschörnig
1
, Robert Wehlitz
1
, Ingo Rößner
1
and Bogdan Franczyk
2,3
1
Institute for Applied Informatics (InfAI), Leipzig University, Hainstr. 11, 04109 Leipzig, Germany
2
Information Systems Institute, Leipzig University, Grimmaische Str. 12, 04109 Leipzig, Germany
3
Business Informatics Institute, Wrocław University of Economics, ul. Komandorska 118-120, 53-345 Wrocław, Poland
Keywords: Internet of Things, Value-added Services, Platform Architecture.
Abstract: The Internet of Things (IoT) is based on ubiquitous smart devices equipped with sensors, actuators and tags
which are connected to the Internet. Combining their sensing and actuation capabilities yields large
possibilities for businesses to provide customers with value-added services in terms of energy management,
entertainment, security and convenience. Businesses may use IoT data to develop innovative business models
thus further increasing the value and usefulness of smart devices. Despite these potentials, providers of IoT
platforms struggle to tackle some challenges which come with the design and operation of the platforms. In
this paper, we propose an architectural concept which aims to bridge this gap and provides an integrated
environment for smart device integration, data and analytics as well as IoT-aware processes. We also present
the Smart Energy Platform (SEPL) as an instantiation of the proposed concept to evaluate it in terms of its
functional feasibility and show that it addresses the issues current IoT platforms face.
1 INTRODUCTION
The Internet of Things (IoT) comprises smart devices
equipped with sensors, actuators, and tags which are
or attached to real-world objects and connected to the
Internet (Han et al., 2016). The number of smart
devices is estimated to reach 24 billion by the year
2020 (Greenough, 2016). This increase is
accompanied by huge amounts of data which offer
new possibilities for businesses to provide
value-added services for their customers.
In the energy domain, current and future trends,
such as smart metering, smart home, e-mobility and
smart grid are summarized under the umbrella term
smart energy (Aichele et al., 2013). In this context,
smart devices, which provide sensing and actuation
capabilities, are already used to gather energy data
and respond to changing circumstances, e.g. in the
case of demand-side management of energy
consumption. IoT platforms are a means to
interconnect distributed smart devices, use their
exposed services to create more sophisticated
services with added value and to build applications
upon them. Against this background, previous
research has found that current IoT platforms are
lacking important features for device integration, data
and analytics, as well as IoT-aware processes
(Wehlitz et al., 2017).
In this light, the main contribution of this paper is
an architectural concept for an IoT platform, which
bridges this gap. We furthermore present a software
prototype of a Smart Energy Platform (SEPL) as an
instantiation of the proposed concept in order to
evaluate its functional feasibility.
In this paper, we describe the motivation for
conducting this research (Sect. 2). Following, we
highlight the major challenges and state of the art for
designing and operating IoT platforms (Sect. 3). In
Sect. 4, we introduce our architectural concept, which
is divided into the three technical layers: Integration,
data and analytics, and IoT-aware processes.
Afterwards, we present the SEPL software prototype
to show in which way the components of each layer
are implemented and how they address the identified
challenges (Sect. 5). Finally, the paper concludes with
a short summary and outlook (Sect. 6).
Zschörnig, T., Wehlitz, R., Rößner, I. and Franczyk, B.
SEPL: An IoT Platform for Value-added Services in the Energy Domain.
DOI: 10.5220/0006695205930600
In Proceedings of the 20th International Conference on Enterprise Information Systems (ICEIS 2018), pages 593-600
ISBN: 978-989-758-298-1
Copyright
c
2019 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
593
2 MOTIVATION
The growing number of smart devices offers many
benefits for people in different areas of their everyday
life, such as energy management, entertainment,
security, and convenience (Zion Market Research,
2017). However, smart devices, in comparison to
regular devices, do not offer a notable advantage in
possibilities if the provided data and actuation
capabilities are not utilized. Therefore, services
which use the aforementioned capabilities are
required in order to create customer value around
products like smart meters, smart thermostats, smart
lights, etc.
Especially in the energy domain, utilities have to
develop new service offers and business models due
to low energy prices and high dues. In this context,
expert interviews revealed that utilities’ margin for
electricity supply in Germany is between 10 and 20
euros per household per year. Hence, with the
comprehensive rollout of smart meters and the
increasing adoption of smart devices, such businesses
have the chance to provide services which, in addition
to the pure energy supply, add value to their
customers and enable new sources of income.
Possible fields of application could be, but should not
be limited to, the increase of energy consumption
transparency, energy savings, and energy efficiency
in households.
Against this background, we found that IoT
platforms are an important building block to provide
the infrastructure and software tools necessary to
develop and run value-added services in the energy
domain. They can be used to integrate smart devices
of customers, e.g. smart meters, process and analyse
sensing data, e.g. energy consumption values, and
create more sophisticated services by service
composition and individualisation, for instance, to
control household appliances in an energy efficient
manner.
3 STATE OF THE ART
Recent studies analyzing IoT platforms have
identified multiple challenges for their design and
operation. These are usually separated into a
functional and non-functional dimension (Wehlitz et
al., 2017).
On the side of the functional challenges, data
management, service abstraction, marketplace
functionality, and device and service discovery are
most significant.
The area of non-functional challenges for IoT
platforms contains interoperability, privacy and
security, scalability, quality of service and context
awareness.
With regard to these challenges, there are many
existing approaches from businesses, the open source
community as well as researchers which have
different degrees of coverage (Wehlitz et al., 2017).
Concerning the functional challenges, most platforms
use service abstraction for providing a unified access
to smart devices. Data management and device and
service discovery are implemented only partly by
some platforms. More importantly, a marketplace is
missing in almost all of them.
On the side of the non-functional challenges,
context awareness is implemented by almost all
platforms. Yet, the integration of smart devices of
different vendors with distinct service interfaces, i.e.
interoperability, is supported only in part. Also, the
bulk of the regarded platforms do not fully support
scalability which seems to be a major issue with rising
smart device numbers. Quality of service and privacy
and security are also challenges which are only partly
solved (Wehlitz et al., 2017).
It is noteworthy, that two platforms we have
studied, FIWARE and ThingWorx, address most of
the found challenges. However, both are extensive
systems, exhibiting high complexity, a steep learning
curve or they are proprietary. Hence, it seems
plausible that they are not suitable to be used directly
by customers.
4 PLATFORM ARCHITECTURE
Looking at the challenges and the state of the art as
described in Sect. 3, we found the basic design of an
IoT platform should comprise three layers:
An integration layer providing bidirectional
platform-to-device communication as well as
tools to unify device and service access.
A data and analytics layer providing smart
device data management, but also means to
analyze and filter data.
An IoT-aware process layer to enable the
composition and orchestration of IoT and
analytics services into more useful services and
applications.
Additionally, a marketplace should be offered to
platform users allowing them to share created content
in order to bring the platform to life. Another
important design goal for us is to make the entire
platform loosely coupled and easily extensible. This
requires
a unified, resilient way of communication
ICEIS 2018 - 20th International Conference on Enterprise Information Systems
594
Figure 1: Components of the integration layer and their interactions.
between the different system components. The
proposed concept employs a streaming platform to
achieve this.
4.1 Integration Layer
The integration layer is the foundation of the
proposed platform concept. It acts as the main point
of integration for different kinds of smart devices.
Therefore, the layer must provide interfaces for them
to send and receive data.
In the context of smart home and smart building
environments, smart devices are often connected to a
local gateway (Risteska Stojkoska and Trivodaliev,
2017). In this regard, the platform operator needs to
provide either best practices for end users on how to
configure their gateway device or, more commonly,
find ways to enable inbound network traffic
automatically. In this regard, they have to provide
software for different gateway types in order to
register locally connected smart devices on IoT
platforms.
As shown in
Figure 1, we use platform connectors
in conjunction with client connectors to achieve this.
Client connectors are deployed either on gateways (1)
or on stand-alone smart devices (2). Supporting
various message protocols (e.g. HTTP, WebSocket,
CoAP), they send data to their counterparts, i.e.
platform connectors. These receive the data and push
it to the central streaming platform (3) for further
processing by other IoT platform services.
Publish-subscribe communication (e.g. based on
MQTT) (4) via an intermediary broker (5) is also
possible. Besides the actual communication with
smart devices, platform connectors gather historical
data about connection states of client connectors, e.g.
time of disconnect. They also handle the outgoing
data, meaning service requests, to the client
connectors. These requests are pushed to the
streaming platform by IoT platform services and
applications. The platform connectors interface the
streaming platform, fetch new service requests (6),
and proceed in sending the data to the destined client
connector. In addition, platform connectors register,
unregister and update device and service instance data
in the device and service repository (7). It contains
device and service type descriptions as well as device
instances with their service instantiations.
Since the IoT domain has not yet undergone
successful standardization efforts (Díaz et al., 2016),
it is characterized by a large heterogeneity in terms of
data types, semantics and protocols, which
complicates access to their services (Perumal et al.,
2015; Díaz et al., 2016). Therefore, it seems
appropriate to transform different data types and
structures as well as semantics to a unified basis. In
our proposed concept, this is done by the platform
connectors. They map data structures of smart
devices to a generic data model, thus enabling
platform wide understandability and reusability
across different services and applications.
In this model, a device type is a collection of
comparable devices (e.g. Temperature Sensor),
which contains services (e.g. getTemperature).
Transforming data structures of smart devices into the
generic data model ensures that all data, which is
queued in the streaming platform, is already unified.
This approach also leads to reusability of value types,
since they are used frequently to create services in the
platform environment. Also, data structures using this
data model are independent from specific
serialization
formats. Hence, the layers on top of the
integration layer are able to create processes and
applications with high reusability across smart
devices.
IntegrationLayerIntegrationLayer
Platform
Connector
Platform
Connector
Deviceand
ServiceRepository
Deviceand
ServiceRepository
Streami ng
Platform
Streami ng
Platform
EdgeDe vi ceEdgeDe vi ce
BrokerBroker
Edge/MobileDe vi ceEdge/MobileDe vice
Cli entConnectorCli entConnector
EdgeDe vi ceEdgeDe vice
GatewayGateway
Cli entConnectorClientConnector
1
2
4
5
3
7
6
SEPL: An IoT Platform for Value-added Services in the Energy Domain
595
Figure 2: Components of the data & analytics layer and their interactions.
4.2 Data & Analytics Layer
The data and analytics layer is on top of the
integration layer and provides sophisticated data
manipulation, analytics, persistence and event
processing functionalities. Smart device data usually
arrives in data streams and is characterized as time
series data, thus creating the need for real-time
processing (Pawar and Attar, 2016). While the batch
processing of historical data combined with the
application of prediction models on this data is
already established, real-time processing of smart
device data enables the detection of causable
relationships between the devices themselves and
their environments. This may lead to the autonomous
adaption to new situations (Stolpe et al., 2016). As a
result, the analytics layer should be a lightweight
stream processing system which is able to easily
access the central streaming platform with the smart
device data, as shown in
Figure 2. Also, it is necessary
to provide capabilities to design analytics flows for
large amounts of different analytics problems.
In order to process data, we utilize analytics
operators which perform analytics actions on data
streams (e.g. Temperature Unit Conversion). These
operators are lightweight and perform a single task.
They may be composed into complex analytics flows,
thus providing the needed flexibility in analytics
design. Analytics operator metadata is stored in the
analytics operator repository and comprises inputs
(e.g. Temperature #1: float, Unit: string), outputs (e.g.
ConvertedTemperature: float), and configuration
parameters (e.g. ConversionUnit: string).
A flow designer is used to graphically design and
orchestrate input streams and analytics operators into
analytics flows, using operator metadata (1). Finished
analytics flows (e.g. Temperature Unit Conversion
and Mean Value) are saved into the flow repository
(2) and may contain multiple analytics operators and
data streams as input. A flow exposes inputs (e.g.
Temperature #1: float, Unit #1: string, Temperature
#2: float, Unit #2: string), outputs (e.g.
MeanTemperature: float) and configuration items
(e.g. ConversionUnit: string) for all analytics
operators used. Analytics flows are started (3) and
monitored (4) using a flow executor. In addition, the
flows are linked to smart device services data (5) as
input streams.
The analytics operators pull the data from the
streaming platform (input topic #1), process it and
push it back (output topic #1) (6). Analytics operators
linked to them in an analytics flow, pull their input
data from the data pushed back (output topic #1) and
continue in the same manner as the previous analytics
operators. In order to achieve data persistence, all
data streams need to be pulled from the streaming
platform into a Data Lake (7). Following the Data
Lake paradigm, they are saved as they are without
further processing. Once the saved data is needed by
platform services, it is pushed back into the streaming
platform.
Another key aspect of the data and analytics layer is
event processing. An event processing engine is
consuming data streams and triggers events if defined
event rules are met (8). We describe events at type
(e.g. value > 22) and instance level (e.g. temperature
value
from Temperature Sensor #1 > 22 °C). Events
are defined using an event rule designer and event
rules are stored in the event rule repository (9). The
event processing engine listens for all event rules
(10). These may be applied to all data streams present
in the streaming platform (e.g. smart device data,
analytics data, process data, etc.). In this regard, the
event rule designer may also access the analytics flow
repository to expose analytics flows to the user which
may be used to listen on for event detection (11).
Streaming
Platform
Streaming
Platform
StreamProcessingSystemStreamProcessingSystem
Analytics
Operator
Analytics
Operator
FlowExecutorFlowExecutor
Analytics
Operator
Reposit ory
Analytics
Operator
Reposit ory
FlowDe si gne rFlowDe si gne r
FlowRepositoryFlowRepository
Even t
ProcessingEngine
Even t
ProcessingEngine
Even tRule
Designer
Even tRule
Designer
EventRule
Reposit ory
EventRule
Reposit ory
Deviceand
ServiceRe pository
Deviceand
ServiceRe pository
1
2
3
7
5
4
6
11
8
9
Event
ProcessingEngine
Event
ProcessingEngine
10
ICEIS 2018 - 20th International Conference on Enterprise Information Systems
596
Figure 3: Components of the IoT-aware process layer and their interactions.
4.3 IoT-Aware Process Layer
The IoT-aware process layer bundles the
functionalities of the integration as well as the data
and analytics layer. It serves the orchestration of IoT
services, analytics and event data along businesses,
smart devices and customers. The layer is based on a
centralized orchestration model and allows for the
modelling, implementation and execution of
IoT-aware processes.
As illustrated in Figure 3, processes are designed
using a process designer, which includes a graphical
user interface and stores created process models in the
process repository (1). It has access to device and
service types from the device and service repository
(2) together with event rule types from the event rule
repository (3). Hence, processes are designed using
abstract device, service and event definitions. This
allows for processes which are instance-independent
and therefore easily reusable. In addition, the same
model can be deployed for different smart devices
multiple times. Executable process models are
deployed by the process deployment component. It
pulls model definitions from the process repository
(4) and scans it for references to device/service and
event types. If any are present, it retrieves compatible
instances from the device and service repository (5).
User input is required to select which concrete
devices and services shall be used for a process
deployment. Event types are handled in the same
manner. After this, the process model is deployed to
the process engine, which ensures proper execution
(6). External task workers execute service tasks. Once
the process engine encounters a service task (e.g.
smart device actuation), it creates a new job for the
external task workers to be pulled and executed.
When they completed their jobs, they notify the
process engine which proceeds to the next step (7). In
case of an actuation service, an external task worker
fetches the required service description from the
device and service repository and sends a service
request to the streaming platform (8), which in turn is
accessed by platform connectors. A platform
connector forwards the service request to the client
connector which executes it or relays it to the actual
smart device.
Sensing requests get relayed back from the device
to the client connector to the platform connector
which pushes the sensing data to the streaming
platform. From there, events are, if applicable,
applied by the event processing engine on the data
stream. The process engine is able to process fired
events and act according to the executed process
model (e.g. to start a new or to affect the control flow
of a running process instance). The process engine is
supervised by the process monitoring component,
which provides insights to historical and operational
data of process instances (9). Finally, the execution of
processes may be integrated in external applications
using a REST API (e.g. a mobile application to trigger
new process instances) (10).
4.4 Marketplace
The marketplace allows customers to share their data,
modeled IoT-aware processes, analytics flows and
operators with other platform users (businesses and
customers). All content created by customers and
businesses as well as data streams originating from
smart devices may be accessed as defined in the
Proc essEngineProcessEngine
Process
Repository
Process
Repository
Proces s
Designer
Proces s
Designer
External
Tas kWorker
External
Tas kWorker
Proces s
Monitoring
Proces s
Monitoring
Event
Processing
Engine
Event
Processing
Engine
Device
andService
Repository
Device
andService
Repository
Streaming
Plat form
Streaming
Plat form
Proces s
Deployment
Proces s
Deployment
Ret rieveDeviceand
ServiceTy pes
Ret rieveDeviceand
ServiceTy pes
Ret rieveDevice
andService
Instances
Ret rieveDevice
andService
Instances
Fetch/CompleteTa sksFetch/CompleteTa sks
Ret rieve
Process
InstanceData
Ret rieve
Process
InstanceData
Apply
Even tRules
Apply
Even tRules
Deploy
ProcessModel s
Deploy
ProcessModel s
Save/Load
ProcessModels
Save/Load
ProcessModels
Ret rieve
ProcessModels
Ret rieve
ProcessModels
Inv ok eServicesInv ok eServices
TriggerEventsTriggerEvents
Third
pa rty
IoTApplication
Third
pa rty
IoTApplication
External
Interface
Event
Rule
Repository
Event
Rule
Repository
StartEventRule
Instances
StartEventRule
Instances
Ret rieveEv entRu leTypesRet rieveEv entRu leTypes
1
2
3
4
Ret rieveEv ent
RuleInstances
Ret rieveEv ent
RuleInstances
5
6
7
Send
Req uests
Send
Req uests
8
10
9
Event
RuleDesigner
Event
RuleDesigner
Save/Load
EventRuleTy pes
Save/Load
EventRuleTy pes
SEPL: An IoT Platform for Value-added Services in the Energy Domain
597
permission system. This system links user instances
to processes, flows, operators, and device instances
from the respective repositories. Customers and
businesses are able to share their content and device
data and publish them at the marketplace. The access
to the shared entities may be limited to individual
customers or businesses or groups of them. Also, a
fee for using the entities may be applied. If other
customers or businesses decide to subscribe to
content or device instances, the required permissions
are set in the permission system.
Granting permission to content allows the
subscriber to use the flow, operator or process with
their own smart devices. In case of a shared smart
device, the subscriber may access emitted data from
the device as set by the owner. In this context, using
the actuation capabilities of a shared device is
generally forbidden because of security concerns.
5 PROTOTYPE
In order to provide a proof of concept of our
architecture presented in Sect. 4, we implemented all
system components as an integrated cloud-based
software prototype, the SEPL. Its purpose is to
provide platform tools to customers, to integrate their
smart devices, analyze their data and develop and use
IoT-aware processes for energy management
scenarios. Also, customers may share their data with
businesses, which in turn are provided with tools to
develop value-added services around this data,
creating surplus value.
The platform architecture is based on the concept
of Microservices, meaning that functionalities are
encapsulated as single services by software
containers, in our case using Docker
(https://www.docker.com/). This provides easy
scalability of services which undergo heavy usage
and enables easy change or substitution of services
with changed requirements. Because of this, we
employed Apache Kafka as a central streaming
platform (https://kafka.apache.org/). To ensure
usability, we also developed a front-end application
using AngularJS which provides access to important
platform tools. User authentication and authorization
is ensured using the API gateway Kong
(https://getkong.org/) in conjunction with a user
repository and permission system. In order to further
strengthen privacy and security we followed the
“Privacy by Design” and security principles as
pointed out in scientific literature (Schaar, 2010;
Dougherty et al., 2009).
Because of the complex structure and the
Microservice approach of the platform architecture, it
is necessary to employ a powerful container
orchestration and scheduling service. We use Rancher
(http://rancher.com/) as it is easy to use and offers
many out-of-the box features for platform operation.
5.1 Integration Layer
The main focus when implementing the integration
layer, was to provide interfaces to enable
communication between smart devices and the SEPL,
but also to unify device and service descriptions from
different vendors. Although the layer components are
designed to be easily extendable in terms of protocol
and device integration, a lot of the smart devices we
use in our laboratory environment for testing our
approach support the wireless communication
standards ZigBee and Z-Wave. Both are very
common and well suited for home energy
management scenarios (Zhao et al., 2016).
The client connectors in smart home and smart
building environments are provided in two ways: As
plugins for already existing IoT gateways and as
software libraries to be used for single smart devices
by more advanced users. When deployed at IoT
gateways, it is possible to use existing resources of
them to search for locally connected smart devices
and send this information to the SEPL, as the
foundation for their integration.
The platform connector is scalable in order to
being able to communicate with huge numbers of
smart devices. It uses the device and service
descriptions from the device and service repository to
transform data from client connectors to the generic
data model as explained in Section 4.1.
The device and service repository is based on
Virtuoso (https://virtuoso.openlinksw.com/) and
utilizes RDF for schema description. Other services
can access the data from the device and service
repository using a REST interface. Since it also stores
all registered device instances, it enables platform
services and users to discover smart devices and their
services. Internally, it queries the RDF database using
SPARQL, therefore enabling powerful queries.
5.2 Data & Analytics Layer
The analytics concept we employ is based on the
Kappa architecture, meaning all data is handled as a
stream. Therefore, analytics operators use the Kafka
Streams (https://kafka.apache.org/) library to
integrate with the streaming platform and are written
as individual Microservices, which are encapsulated
ICEIS 2018 - 20th International Conference on Enterprise Information Systems
598
using Docker. Using JAVA or Python, it is possible
to create powerful data analytics as well as machine
learning algorithms for processing. This allows for
flexible analytics operators which offer the same
capabilities as stream processing engines like Apache
Spark Streaming or Apache Storm. In addition, this
approach helps in reducing the programming
overhead caused by changing data models and
analytics requirements, compared to a Lambda
architecture (Zschörnig et al., 2017).
New analytics flows are created with the help of
the analytics designer, which interfaces the analytics
operator repository. In our prototype, the designer
utilizes a flow chart visualization to enable users to
design new analytics flows, which are saved in the
flow repository.
The request to start an analytics flow is send to the
flow executor, which checks it for errors. If the flow
is valid, it starts the analytics operator Docker
containers in the required order with the necessary
input, output and configuration values. The analytics
operator containers are deployed on the underlying
orchestration and scheduling environment. Data
persistence is achieved by pulling all data streams
from Apache Kafka into the Data Lake which uses a
HDFS environment. Platform services may request
data from the Data Lake, which pushes it back into
Apache Kafka.
Event rules are stored in the event rule repository
and are created using a JSON document which
contains information on the data stream to be
monitored and the actual rule. A rule is written by
combining service fields from the device and service
repository and logical expressions using the event
rule designer.
5.3 IoT-Aware Process Layer
The process designer component of our software
prototype is based on bpmn.io, a BPMN 2.0 rendering
toolkit and web modeler (https://bpmn.io), which was
extended to integrate with the device and service
repository. We selected BPMN 2.0 because it seems
to be the best suited modeling language for mapping
IoT concepts in processes (Meyer et al., 2011).
Following Meyer et al., 2013, our solution intends
that IoT-aware processes are modeled within BPMN
pools. The lanes of a pool represent device types.
Each service task that is placed on a lane is assigned
to and handled by the respective device type.
In case of a new deployment, the process
deployment component analyses the process
definition and prompts the user to select concrete
instances for every found event and/or device type. It
then deploys the process model to the process engine,
which is based on the open-source platform Camunda
(https://camunda.org). This supports the external task
pattern and, hence, provides more flexibility and
scalability for executing service tasks. The
implementation of external task workers is
independent from a specific programing language and
additional source code for service task execution has
not to be deployed to and executed by the process
engine. Additionally, in conjunction with the event
processing engine and the stream processing system,
the process engine enables context awareness. In this
regard, the stream processing system provides new
insights into smart device data and the event
processing engine is able to define actions based on
changing parameters. IoT-aware processes may
include this information, thus allowing smart devices
to be “aware” of their surroundings and adapt to
changing environments.
5.4 Marketplace
The main objective of the marketplace is to allow
users (customers and businesses) of the SEPL to
easily share and subscribe to processes, analytics
operators and flows. In this regard, users of the SEPL
may define if a process, analytics flow or operator
may be published at the marketplace or only be shared
with a single user or a group of users. If it is published
at the marketplace, a request is sent to its repository,
which then lists it in the appropriate category. This
categorization is based on the type of the shared entity
(e.g. process vs. operator) and its characteristics (e.g.
inputs, outputs, purpose, etc.).
In conclusion, the marketplace offers various
capabilities to share key entities of the SEPL, thus
enabling device integration through sharing, but most
important it is the key component to enable
businesses to apply new business models around
smart devices in the energy domain.
6 CONCLUSION & OUTLOOK
In this paper, we presented an architectural concept
for the design and operation of IoT platforms.
Customers may use its tools to integrate and
orchestrate their smart devices and gain insights into
their data, whereas businesses use them to create new
applications and value-added services on top of the
customer-specific data.
The proposed concept takes into account
functional and non-functional challenges which were
identified by current research and are not met entirely
SEPL: An IoT Platform for Value-added Services in the Energy Domain
599
by existing solutions. Looking at these challenges, we
found that an IoT platform should comprise technical
layers for device and service integration and
abstraction as well as their discovery, but also
capabilities for data management and analytics and
IoT-aware process modeling, implementation and
execution. Furthermore, customers and businesses
need to be able to share device data access and
processes with other users of the platform through a
marketplace. The composition of all these platform
components leads to quality of service regarding the
communication with smart devices and context
awareness in terms of their environment.
Future research in this field needs to focus on the
advanced use of semantic technologies for device
integration as well as device and service discovery.
Analytics architectures have to be developed to cater
to a non-technical audience, allowing self-service
analytics. In terms of IoT-aware processes, tools for
adaptive case management should be investigated to
improve context awareness and flexibility. Also,
business models on how to operate this kind of
platform need to be created and evaluated. Finally,
the research concerning hybrid approaches on IoT
platforms, e.g. in conjunction with Fog Computing,
needs to be deepened.
ACKNOWLEDGEMENTS
The work presented in this paper is partly funded by
the European Regional Development Fund (ERDF)
and the Free State of Saxony (Sächsische Aufbaubank
— SAB)
REFERENCES
Aichele, C., & Doleski, O. D. (Eds.). (2013). Smart Meter
Rollout - Praxisleitfaden zur Ausbringung intelligenter
Zähler. Wiesbaden: Springer Vieweg.
Díaz, M., Martín, C., & Rubio, B. (2016). State-of-the-art,
challenges, and open issues in the integration of Internet
of things and cloud computing. Journal of Network and
Computer Applications, 67, (pp. 99–117).
Dougherty, C., Sayre, K., Seacord, R. C., Svoboda, D., &
Togashi, K. (2009). Secure Design Patterns. Retrieved
from http://www.dtic.mil/get-tr-doc/pdf?AD=ADA636
498
Greenough, J. (2016). How the 'Internet of Things' will
impact consumers, businesses, and governments in
2016 and beyond. Retrieved from
http://www.businessinsider.com/how-the-internet-of-
things-market-will-grow-2014-10?IR=
Han, S. N., Khan, I., Lee, G. M., Crespi, N., & Glitho, R.
H. (2016). Service composition for IP smart object
using realtime Web protocols: Concept and research
challenges. Computer Standards & Interfaces, 43, (pp.
79–90).
Meyer, S., Ruppen, A., & Magerkurth, C. (2013). Internet
of Things-Aware Process Modeling: Integrating IoT
Devices as Business Process Resources. In D.
Hutchison, T. Kanade, Ó. Pastor (Eds.), Lecture Notes
in Computer Science. Advanced Information Systems
Engineering (Vol. 7908, pp. 84–98).
Meyer, S., Sperner, K., Magerkurth, C., & Pasquier, J.
(2011). Towards modeling real-world aware business
processes. In D. Guinard, V. Trifa, & E. Wilde (Eds.),
In: The Second International Workshop on Web of
Things (p. 1).
Pawar, K., & Attar, V. (2016). A survey on Data Analytic
Platforms for Internet of Things. In CAST-2016. 19-21
December 2016 (pp. 605–610). Piscataway, NJ: IEEE.
Perumal, T., Datta, S. K., & Bonnet, C. (2015). IoT device
management framework for smart home scenarios. In
2015 IEEE 4th Global Conference on Consumer
Electronics (GCCE) (pp. 54–55).
Risteska Stojkoska, B. L., & Trivodaliev, K. V. (2017). A
review of Internet of Things for smart home:
Challenges and solutions. Journal of Cleaner
Production, 140, Part 3, (pp. 1454–1464).
Schaar, P. (2010). Privacy by Design. Identity in the
Information Society, 3(2), (pp. 267–274).
Stolpe, M. (2016). The Internet of Things: Opportunities
and Challenges for Distributed Data Analysis. ACM
SIGKDD Explorations Newsletter, 18(1), (pp. 15–34).
Wehlitz, R., Häberlein, D., Zschörnig, T., & Franczyk, B.
(2017). A Smart Energy Platform for the Internet of
Things-Motivation, Challenges, and Solution Proposal.
In Business Information Systems: 20th International
Conference, BIS 2017, Poznan, Poland, June 28-30,
2017, Proceedings (Vol. 288, p. 271).
Zhao, Z., Agbossou, K., & Cardenas, A. (2016).
Connectivity for Home Energy Management
applications. In APPEEC 2016. 2016 IEEE PES Asia
Pacific Power and Energy Engineering Conference:
October 25-28, 2016, Xi'an, China (pp. 2175–2180).
Zion Market Research. (2017). Global Smart Home Market
is Set for a Rapid Growth and is Expected to Reach
around USD 53.45 Billion by 2022. Retrieved from
https://www.zionmarketresearch.com/news/smart-
home-market
Zschörnig, T., Wehlitz, R., & Franczyk, B. (2017). A
Personal Analytics Platform for the Internet of Things:
Implementing Kappa Architecture with Microservice-
based Stream Processing. In Proceedings of the 19th
International Conference on Enterprise Information
Systems (pp. 733–738).
ICEIS 2018 - 20th International Conference on Enterprise Information Systems
600