Article Written by Ingo Kiesewetter, Vice
President Enterprise Architecture at itelligence, Inc.
Everybody is talking about IoT as
the best things since sliced bread. What
often gets neglected in all this talk about the potential of IoT is how to make
it work. IoT presents unique challenges
to common system architectures. Platform-as-a-Service
(PaaS) is an important layer in the stack.
We will lift the curtain back and peek at what is behind.
But first, let's clear up any
potential confusion: IoT describes physical
objects like machines, appliances and sensors that are connected using the
IoT as enabler for business transformation
IoT opens the door to transformational
changes and new business models. It allows companies to reinvent themselves by
removing technical constraints that exist today. Today, a typical manufacturing
company might make equipment and then offer repair and maintenance services. With IoT the business model can change to
sharing the added value of additional equipment uptime. Sensors in the equipment can relay relevant
metrics to the manufacturer's backend systems and trigger service tickets
before an issue leads to equipment downtime.
The availability of IoT also
offers opportunities for completely new business models. One example is an e-mobility company that
uses IoT sensors in charging stations and cars to provide customers with a
real-time view of available cars and the status of charging stations.
Connecting large numbers of IoT
devices to back office systems poses specific challenges. The technology forms the backbone for the
transformational changes to business models but is often overlooked.
IoT devices need to be connected to a central
layer that integrates them into backoffice processes. The number and geographic distribution of IoT
devices poses unique challenges around device management and security. The data volumes generated by IoT sensors can
be vast. Commonly these issues are
addressed by using edge
of fog computing. These terms describe a cloud-like, often
distributed and decentral layer that isolates IoT communication from the rest of
the overall process in business systems.
In addition to connecting devices the data being generated can be
aggregated, filtered and analyzed in this layer. Security is another layer than needs to be carefully
considered in IoT projects. Decoupling
devices in the open Internet from protected business systems is a must.
Cloud Technology Partners
Many providers offer products for
device connectivity. Let's look at SAP's
Hana Cloud Platform (HCP) as an
example: HCP is fundamentally a Platform
as a service (PaaS) platform. SAP has
developed numerous packages in HCP including commonly used IoT services (device
connectivity and management, communication, data streaming, security, SQL
Anywhere etc.). This means that HCP can
act as an edge computing platform as well as a PaaS platform for connectivity
to backend systems.
Connecting IoT to Business Processes
Once data generated by IoT
devices is available in a digestible format (aggregated, filtered etc.) it can
be used in the context of a business process in the core back office systems
like SAP S/4HANA. The integration is
commonly built in a cloud-based platform (PaaS) or an on-premise integration
tool. By separating the IoT
device-facing activities from the more sensitive backoffice-facing ones, possible
performance and security issues can be mitigated more easily. Since IoT scenarios often change more often
and more quickly than back office processes, using a PaaS platform allows
changes to the IoT scenario without impacting the stable core process.
Using PaaS to integrate 3rd
party applications into a business process in a back office system has been a
common approach for several years. By
using defined APIs in the back office system, a stable interface layer can be
defined, allowing for a more agile approach in the content developed and
deployed in the PaaS. With emerging
technologies like IoT, this added agility is particularly important as
requirements are often not clearly defined and morph as an implementation
project moves along. ‘Classic'
on-premise integration tools can be used in a similar manner, but most
providers of such tools are slower to roll out new features and capabilities
compared to PaaS providers. To
illustrate, compare the features in SAP NetWeaver Process Orchestration (PO)
with SAP's PaaS solution (Hana Cloud Platform):
while both tools have the mapping and orchestration capabilities one can
expect from an integration tool, SAP made a number of IoT-specific services
available only in HCP. Some of these
services are more device-oriented (as described above) but some like data
streaming can be very useful in integrating IoT devices with a back office
When planning an IoT project one
needs to consider the technical landscape carefully. Since communication between and from devices
is executed over the open Internet a layer to isolate devices from the rest of
the landscape is needed. Fog or edge
computing solutions are commonly used for this purpose. Integration to a back office system is
(mostly) necessary to realize the value IoT promises. PaaS platforms often support this better than
classic on-premise integration tools.
About the Author
Ingo Kiesewetter is the Vice
President Enterprise Architecture at itelligence, Inc. in Cincinnati, OH. He has over 21 years of experience in
implementing business software, designing and building complex integration
scenarios and guiding clients through the process of transforming their
business through IT.
A native of Germany he is now a
US citizen and resides in Cincinnati, Ohio with his wife, two dogs and four cats.