federated infrastructure, in which Virtual Entities are clustered based on
similarity. Dedicated places are in charge of the IoT Services they offer and
provide their descriptions as part of a distributed resolution framework. The
framework is scalable and fault tolerant because of distribution.
Reduce computational complexity
Whenever possible the system can reduce the computational complexity, thus
leading to a simpler system which needs less time and often energy. As an
example, instead of a complex intrusion detection system, there could either be
no intrusion detection at all (DC PS.13) or a less complex security by design
(DC PS.14), e.g. a protocol stack with built in threshold-based protection
against too many session initiations.
Distribute processing over time
To reduce the number of resources needed it is often possible to distribute
some processing tasks over time, when their results are not immediately
necessary (DC PS.16). In case of hard real-time constrains this might not be
always possible, but many system do not need real-time at all, or do only have
soft real-time constraints. Distributing processing over time can help preventing
the system from scaling or reduce the use of remote (over the web) services.
Minimize used of shared resources
In many IoT systems the most scare and most expensive resource is
bandwidth, especially in wireless battery powered systems. It is necessary to
design the functional components accordingly and especially plan the
deployment to avoid bottlenecks on the devices/resources.
Reuse resources and results
To be able to reuse resources and results the functional components need to be
aware of a history for reuse (DC PS.21). The information model needs be aware
of such caching mechanisms (DC PS.22). In terms of deployment the history
can either be stored locally (DC PS.23), remotely (DC PS.24) or a combination
of both (DC PS.25).
If the information history is stored locally (DC PS.23) the information history is
stored on the IoT device that has produced the information over time. History
information needs to be secured in the same way as the present information to
avoid information leaks. If constrained IoT devices are used, then the storage
size of information history as well as the information processing performance is
limited: Having a local storage place for history information on each IoT Device
requires less device performance and less effort to secure the history, but the
single information host is against good scalability. The availability of information
history depends on the availability of the IoT device hosting the history.
DC PS.24 describes the case, where the information history is not stored on the
IoT Device that has produced the information, but on a different IoT Resource,
to which the information is uploaded to. The additional history resource needs to