Public auditing for real-time medical sensor data in cloud-assisted HealthIIoT system

Weiping Ye, Jia Wang, Hui Tian, Hanyu QuanFrontiers of Optoelectronics 2022 · DOI

Abstract

With the advancement of the Industrial Internet of Things (IIoT), wireless medical sensor networks (WMSNs) have become integral to modern healthcare, enabling the collection of real-time patient data—a field known as HealthIIoT. Due to the limited resources of wearable devices, this data is typically outsourced to cloud servers. However, ensuring data integrity is critical for accurate diagnosis. This paper proposes PAMSD, an efficient public auditing scheme designed for HealthIIoT. By utilizing a novel online/offline tag generation algorithm and a secure hash function for data blinding, PAMSD addresses the dual challenges of real-time processing and patient privacy. Experimental results demonstrate that PAMSD significantly improves tag generation efficiency and auditing performance compared to existing state-of-the-art schemes.

Introduction

In the era of digital health, wearable sensors continuously monitor vital signs such as heart rate and ECG. These lightweight devices lack the storage and computational power to manage the vast amounts of data they generate, necessitating the use of cloud-assisted storage. However, outsourcing introduces security risks, primarily concerning data integrity. If medical data is corrupted or altered in the cloud, it could lead to life-threatening misdiagnoses.

Two primary obstacles hinder effective auditing in this context:

  1. The Real-Time Bottleneck: Traditional auditing requires sensors to perform complex mathematical operations to generate security "tags" for every data packet. This process is too energy-intensive for battery-powered wearables.
  2. The Privacy Paradox: While a Third-Party Auditor (TPA) is needed to verify the cloud's integrity, medical data is highly sensitive. The TPA must be able to verify the data's existence and correctness without ever accessing the actual medical values.

System Architecture

The proposed system involves four main entities: the patient (equipped with sensors), the Cloud Service Provider (CSP), the Third-Party Auditor (TPA), and the medical professional (doctor).

Figure 1: System architecture of the cloud-assisted HealthIIoT auditing framework.

As shown in the architecture, medical data flows from the patient's mobile terminal to the cloud. The TPA manages the integrity verification by issuing challenges to the CSP, which responds with proofs. The doctor can then access the verified data for clinical decision-making.

The PAMSD Scheme

The core of the PAMSD (Public Auditing for Medical Sensor Data) scheme is built upon the Zhang–Safavi–Susilo (ZSS) short signature and bilinear pairings. To optimize performance, the scheme introduces a two-phase tag generation process.

Table 1: Cryptographic notations used in the PAMSD protocol.

Online/Offline Tag Generation

The "heavy lifting" of cryptographic computation is shifted to an Offline Phase. During periods of sensor inactivity, the device pre-calculates random numbers ($w_i$) and auxiliary parameters ($r_i$). When real-time medical data ($m_i$) is captured, the Online Phase requires only a single, lightweight modular arithmetic operation to finalize the tag ($\sigma_i$). This approach ensures that the real-time burden on the sensor is negligible, preserving battery life and reducing latency.

Privacy-Preserving Auditing

To resolve the privacy paradox, PAMSD employs a secure hash function to "blind" the data proof ($\Omega$). Because hash functions are preimage resistant, the TPA can mathematically verify the integrity of the data blocks without being able to reverse-engineer the sensitive medical information. This ensures the TPA remains "credible but curious"—capable of verification but barred from data access.

Performance Evaluation

The researchers evaluated PAMSD against two prominent schemes: CLPA and OOPDP. Testing was conducted using simulated data blocks (4 KB each) ranging from 5,000 to 50,000 units.

  • Efficiency: In the critical online phase, PAMSD's tag generation was significantly faster than CLPA, which lacks an offline/online split.
  • Verification Speed: The TPA's verification time for a standard challenge (e.g., 460 blocks) was approximately 15ms for PAMSD, a dramatic improvement over the 1600ms required by CLPA.
  • Communication Overhead: The proof size remains minimal ($|G_1| + |Z_p^*|$), ensuring that auditing processes do not congest the network.

Conclusion

PAMSD provides a robust solution for the unique constraints of HealthIIoT. By decoupling the computational burden of security tagging and implementing privacy-blinding techniques, the scheme allows for real-time, secure, and private medical data auditing. Future work may explore extending this efficiency to dynamic data operations, such as frequent record updates or deletions, and further reducing the dependency on mobile gateways.

Summary updated

Abstract

With the advancement of industrial internet of things (IIoT), wireless medical sensor networks (WMSNs) have been widely introduced in modern healthcare systems to collect real-time medical data from patients, which is known as HealthIIoT. Considering the limited computing and storage capabilities of lightweight HealthIIoT devices, it is necessary to upload these data to remote cloud servers for storage and maintenance. However, there are still some serious security issues within outsourcing medical sensor data to the cloud. One of the most significant challenges is how to ensure the integrity of these data, which is a prerequisite for providing precise medical diagnosis and treatment. To meet this challenge, we propose a novel and efficient public auditing scheme, which is suitable for cloud-assisted HealthIIoT system. Specifically, to address the contradiction between the high real-time requirement of medical sensor data and the limited computing power of HealthIIoT devices, a new online/offline tag generation algorithm is designed to improve preprocessing efficiency; to protect medical data privacy, a secure hash function is employed to blind the data proof. We formally prove the security of the presented scheme, and evaluate the performance through detailed experimental comparisons with the state-of-the-art ones. The results show that the presented scheme can greatly improve the efficiency of tag generation, while achieving better auditing performance than previous schemes.
BibTeX
@inproceedings{ye2022public,
  author = {Weiping Ye and Jia Wang and Hui Tian and Hanyu Quan},
  title = {{Public auditing for real-time medical sensor data in cloud-assisted HealthIIoT system}},
  booktitle = {Frontiers of Optoelectronics},
  year = {2022},
  doi = {10.1007/s12200-022-00028-1},
}
Figure