Database Reference
In-Depth Information
- Data acquisition from mobile ICOs needs to be flexible to enable pre-processing
and filtering of sensor data on mobile devices while taking into account data
needs expressed by an end application and its users.
- Since sensor data streams can be characterized as Big Data streams, the engine
for continuous processing of such streams needs to be ecient and tailored to
cloud environments to optimally use the available computing resources while
adapting well to the processing load.
- The developed solution should enable delivery of notifications from the cloud
to mobile devices in near real-time and in accordance with user information
needs and context.
- The data acquisition process should be context-aware and quality-driven ,and
influenced by the following parameters: sensor accuracy, mobile ICO location
and battery lifetime, and potential data redundancy/insuciency.
The first three listed objectives can be achieved by use of the cloud-based
publish/subscribe middleware CUPUS, while for the fourth objective we add
an additional component, QoS Manager, capable to perform quality-driven data
acquisition management for mobile ICOs. Naturally, the data produced by mobile
ICOs is geotagged, either by an exact location measured by GPS or cell identifier
(e.g., a mobile network cell identifier or MGRS 2 area). The QoS Manager lever-
ages the publish/subscribe communication style to continuously monitor sensor
locations, their status (e.g. battery level, accuracy) and produced data streams
with the goal to make informed decisions regarding a subset of sensors to keep
active to meet application sensing requirements. Of course, a selection can only
be made for geographic areas with redundant sensors, while the QoS Manager
can also identify areas with insucient sensing coverage. For such areas, the
QoS Manager may employ techniques such as data interpolation and estimation
to meet a required sensing coverage, or even motivate volunteers by means of
incentives to visit and acquire data for those areas.
A view of the OpenIoT architecture for mobile ICOs is given in Fig. 1 .
The central component is the Cloud-based Publish/Subscribe Processing Engine
(CPSP Engine), responsible for acquiring data from external data sources (e.g.,
smartphones), processing the data to see if it matches any active subscriptions,
and disseminating the data to external data consumers. It interacts with a Mobile
Broker (MB), which is a data stream processing component running on mobile
devices responsible for filtering and aggregating locally produced sensor data.
The CPSP engine and MB build the CUPUS middleware. The QoS Manager
interacts with the CPSP Engine to monitor both subscriptions and publications
defined and acquired by the engine to make smart decisions regarding mobile
ICO activation/deactivation. The QoS Manager also serves as a hub for push-
ing the data received by the CPSP Engine to the OpenIoT Cloud Database for
permanent storage. This is achieved by use of the eXtended Global Sensor Net-
work (X-GSN) component which annotates the data and creates RDF triples,
2 The Military Grid Reference System (MGRS) is the geocoordinate standard used
by NATO for locating points on the earth.
Search WWH ::




Custom Search