TABLE I.

STATE-OF-THE-ART MODELS FOR IOT DATA PROCESSING

Proposals	Challenges
[26]	• Used Traditional Cluster management • Data loading efficiency is ignored in the architecture • Only health dataset is evaluated
[27]	• Causes delay in processing • Data accumulation prior to data ingestion is highlighted while data ingestion proficiency is overlook • Classical map-reduce framework used
[28]	• Computation at edge is overlooked during processing
[29]	• The data ingestion proficiency is overlooked and the utilization of MRV1 framework • Limited architecture
[31]	• Conceptual and logical framework • Implementation is not performed
[32]	• It causes additional delay This scheme is also insufficient to load data efficiently to the Hadoop • Conventional cluster resource handling • Only transportation dataset are evaluated
[33]	• Customary and conventional cluster resource handling • It fails to load data efficiently
[34]	• Data loading mechanism exist but delay in loading • Another issue is the utilization of classical map-reduce framework
[35]	• This scheme is also insufficient to load data to the Hadoop • Traditional cluster resource handling is utilized in this scheme
[36]	• Data collection is preferred and data ingestion prior to analysis is overlooked • Only health dataset is evaluated to test the proposed system