.png){kind=icon}
返回
Robust and Lightweight Data Aggregation With Histogram Estimation in Edge-Cloud Systems
.png){kind=icon}
OA
AI
DOI：10.1109/TNSE.2024.3352734.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Secure aggregation based on masked encryption is a crucial technique for data collection in the Internet of Things (IoT) as it employs a lightweight style to enable global data aggregation while protecting individual data. However, network instability makes the design of such schemes more complex as dropout-resiliency is required, where the overheads substantially increase with growing dropped users. Moreover, existing methods primarily concentrate on aggregation and fail to support complex data analysis, such as histogram estimation. This paper proposes a Robust and Lightweight Data Aggregation (RLDA) scheme in edge-cloud systems. RLDA leverages the offline/online paradigm to achieve robust data aggregation, where edge nodes are introduced to assist verifiable key generation offline and data aggregation and key recovery online. RLDA decouples keys of dropped and surviving users so that it can reduce the overhead by always recovering the keys of surviving users rather than reconstructing the keys of growing dropped users. To achieve secure histogram estimation, we design two recoverable aggregation algorithms that support the transformation between vector and single value, and additionally support multidimensional data aggregation. We prove the security and dropout-resiliency of RLDA. The performance shows that RLDA significantly reduces the overhead with growing dropped users.
Keyword:
Encryption
Histograms
Estimation
Data aggregation
Internet of Things
Differential privacy
Privacy
Robust aggregation
Secure histogram estimation
Multidimensional data aggregation
Edge computing
期刊
I
IF:
7.9
论文数:
2.6K
被引数:
10.0K
机构
引用论文
.png){kind=icon}
VeriFL: Communication-Efficient and Fast Verifiable Aggregation for Federated LearningVeriFL: 通信-用于联邦学习的高效且快速的可验证聚合
OAAI
OAAI分享收藏CodedPaddedFL and CodedSecAgg: Straggler Mitigation and Secure Aggregation in Federated LearningOAAI
OAAI分享收藏
查看更多内容