@article{4593,
  author = {Xia Peng},
  title = {A Cloud- based Data Processing Platform for the Internet of Th ings: Architecture, Implementation, and Performance Evaluation},
  journal = {International Journal of Web Applications},
  year = {0000},
  volume = {17},
  number = {4},
  doi = {https://doi.org/10.6025/ijwa/2025/17/4/145-152},
  url = {https://www.dline.info/ijwa/fulltext/v17n4/ijwav17n4_1.pdf},
  abstract = {The paper presents a cloud-based data processing platform for the Internet of Things (IoT), designed to handle
the collection, storage, processing, and visualization of sensor data. It leverages cloud computing to
address the challenges posed by the massive volumes of data generated by IoT devices, utilizing scalable
technologies like Hadoop and Hive for distributed storage and Map Reduce based analytics. The system
architecture includes a TCP server module built with PHP and the Swoole extension to manage real time sensor
data ingestion, and a middleware layer based on Redis to decouple components and enable efficient inter
module communication. The platform offers two primary external interfaces: one for task execution (via
Hive) and another for data uploading to HDFS. Performance tests on the TCP server show high data processing
efficiency at low concurrency, with performance degrading beyond 500 concurrent connections still sufficient
for most real world IoT scenarios where data acquisition rates are lower. The Hadoop cluster demonstrates
robust functionality, including dynamic node addition/removal without service interruption. While domestic
IoT analytics platforms in China lag behind global counterparts, this work contributes a customizable, open
framework that supports secondary development and real time monitoring. The authors conclude that the
platform meets typical production needs and suggest future improvements in API design and developer
usability to lower the barrier for customization and deployment.},
}