Data Life Cycle Management in IoT for Device Data Stores

The data life cycle management in IoT (Abu-Elkheir et al. 2013) for device data stores is shown in Figure 6.3- This figure shows that IoT device data stores start from production to storage in edge devices and from collection to aggregation in front end of communications with other IoT devices. The application/services component deals with query processing for both history and current instances in real time. The back end components deal with pure storage of device data for archiving and updating stored data and performing any pre-processing as required to support for processing and analyzing the data. The complete data life cycle management

Data life cycle management in loT. (Abu-Elkheir et al. 2013.)

Figure 6.3 Data life cycle management in loT. (Abu-Elkheir et al. 2013.)

components are divided in two categories: online (communication-sensitive operations) and offline (storage-sensitive operations).

Communication-Sensitive Operations-. Produce, collect, aggregate, process (in-network), query, filter, update.

Storage Sensitive Operations: archiving, storage/update after processing, pre-processing of stored data.

Query: This component of data life cycle was mostly dealt with the interaction of real-time environment application services where the query was executed and the results were sent for further processing and analyzing for offline components for permanent storage.

Produce-. It is the major component from where this life cycle was initiated for producing data from IoT devices at edge end (in-network) and front end (communications with other IoT devices).

Collect: This collects IoT device data during communication with other IoT devices for integrating and sending their information to further components for summarizing its results.

Filters: This component takes care of deleting the information generated during the collection of data from different IoT devices during integration and tries to send the required information to further components for processing.

Aggregate: This component does summarize all the information from different IoT devices that are integrated for giving the data for back end.

Pre-Processing: It manages many aspects related to missing data, unrelated data, and data cleaning as IoT device data comes in different formats from different sources.

Storage/Update Archiving: This component plays a prominent role in holding data permanently for long-term storage. It takes care of managing IoT device data in defining their type, structure, and other formats of data for easy storage at centralized data store.

Deliver: This component provides the final results to external environment to take necessary decision regarding the output of IoT device data store after getting processed by all components in life cycle.

Process/Analysis: This was a critical component that meant for online processing of IoT device data as it is gathered both in network and in back end. This component does processing of IoT device historical data and predicts future trends in terms of managing its information.

Data Management Framework for IoT Device Data Stores

Figure 6.4 gives the detailed layered architecture for data management framework (Abu-Elkheir et al. 2013) for IoT device data stores. The framework was divided in to six major layers, namely, things layer, communications layer, data layer, federation layer, query layer, and applications/analysis layer. Each layer has its own significance in terms of IoT device data store management.

Things Layer: It consists of actual entities that generate data from the real-time environments and forming groups such as IoT subsystems, mobile and stationary client. This layer consist of in-network query optimizer/executor for generating the query for further processing with related real-time data from the entities and getting the results for responding the system to the clients who are using these entities.

Communications Layer: This layer only uses the latest communication technologies or protocols for transferring the data to the next stacked layers.

Data Layer and Source Layer: Both layers have a critical task to manage the IoT device data for supporting it in many forms and storing the results in persistent data storage. The source layer reports the actual identity of the entities if any new IoT components are added and get notified to data layer to hold the data generated from this layer. The data layers use different data stores for holding metadata, object data, and structured and unstructured data and managing only local IoT devices that are involved in generating the integrated data. The publish/subscribe module and the

Data management framework for loT device data store. (Abu-Elkheir et al. 2013.)

Figure 6.4 Data management framework for loT device data store. (Abu-Elkheir et al. 2013.)

local query analyzer and local data generator of these layers do support in handling any issues related to IoT device data for local IoT system.

Federation Layer: This layer was the center of layered architecture, which tries to manage the integrated global data for IoT subsystem. It creates content profiler for different IoT devices that are involved in generating data and reporting for query layer for further processing. It uses data layer support for holding the local data store and local query optimizer for processing and reporting this layer for getting location aware about other IoT devices and integrating with them for further analysis with global data integrator.

Query Layer: This layer was popular for its major functionality in IoT data management for optimizing queries, executing and referencing them to generate information for the next layer in this framework.

Application/Analysis Layer: This layer was topmost in stack, which actually uses the stored IoT data for further analysis or usage by applications for managing the data and responding to external system users.

 
Source
< Prev   CONTENTS   Source   Next >