Lightweight, Divide-and-Conquer privacy-preserving data aggregation in fog computing

Kinza Sarwar, Sira Yongchareon, Jian Yu, Saeed ur Rehman

Research output: Contribution to journalArticlepeer-review

9 Citations (Scopus)


With the increasing popularity of the Internet of Things’ (IoT) and fog computing paradigm, aggregating IoT data considering privacy concerns over fog networks can be seen as one of the biggest security challenges. Numerous schemes address this problem. However, most of the existing schemes and their associated methods are heavyweight, facing issues related to performance overhead. Furthermore, performing data aggregation at a single aggregator fog node causes an overly computational burden on the node, which results in high latency, degraded reliability and scalability leading to a single point of failure risks. To fill these gaps, this paper presents a lightweight, Divide-and-Conquer privacy-preserving data aggregation scheme in fog computing to improve data privacy, data processing, and storage capabilities. Particularly, we design a data division strategy based on the Level of Privacy (LoP) defined by data owners. The data division strategy not only effectively divides data according to LoP and distributes it among participating fog nodes for aggregation and storage processing, but also reduces computational and memory overhead in the processing simultaneously. Moreover, we perform a privacy analysis of our scheme and perform comprehensive experiments to compare it with other traditional schemes to evaluate performance efficiency. The results demonstrate that our scheme can efficiently achieve data privacy in fog computing and outperforms the other schemes in computational and memory costs.

Original languageEnglish
Pages (from-to)188-199
Number of pages12
JournalFuture Generation Computer Systems
Publication statusPublished - Jun 2021


  • data aggregation
  • IoT security
  • privacy
  • Cloud security


Dive into the research topics of 'Lightweight, Divide-and-Conquer privacy-preserving data aggregation in fog computing'. Together they form a unique fingerprint.

Cite this