Abstract
Cloud computing plays an increasingly vital role in managing the massive data volumes generated by modern hospitals. However, security concerns regarding cloud-stored information persist. This paper proposes a secure cloud storage framework for medical data based on Attribute-Based Encryption (ABE). By designing specific access processes for chief doctors, consulting physicians, and patients, the scheme ensures sensitive information is encrypted locally and accessible only to authorized users. Qualitative security analysis demonstrates the reliability and applicability of this multi-layered defense strategy.
Introduction
Modern medical facilities are "data factories." A mid-sized hospital can generate approximately 60 GB of imaging data—such as CT scans and ultrasounds—every day. Traditional local storage infrastructure often struggles to keep pace, leading to data being moved to offline archives where retrieval is cumbersome for clinicians.
While cloud storage offers the necessary scalability, it introduces significant security risks. Data stored in the cloud is potentially exposed to unauthorized access or untrustworthy service providers. If a breach occurs and data is stored in plaintext, patient privacy is compromised. To address this, a system is needed that leverages cloud scalability while maintaining strict, role-based access control that mirrors the complex hierarchy of a hospital.
Methodology: Attribute-Based Encryption (ABE)
The core of the proposed framework is Attribute-Based Encryption (ABE). Unlike traditional encryption where a specific key unlocks a specific file, ABE allows data to be encrypted based on a set of descriptive attributes.
Granular Access Control
The system defines access strategies using logic gates (AND, OR). For instance, a medical record might only be decryptable by a user who possesses attributes matching a specific department within a specific hospital.
Key technical components include:
- Local Encryption: Sensitive data is encrypted on the hospital's local side before upload. The cloud server only stores and manages ciphertext.
- CA/Cloud Separation: A Certificate Authority (CA) issues keys based on attributes, while the Cloud Server stores data and validates IDs. This ensures the entity holding the data never has the power to generate the keys to read it.
- Blacklist Mechanism: To handle staff turnover or changes in authorization, the cloud server maintains a real-time blacklist. Even if a user has the correct attributes, access is denied if they appear on this list.
System Workflow and User Roles
The framework defines specific protocols for three primary user types, ensuring that data access is tailored to clinical needs and privacy requirements.
- Chief Doctor: Acts as the data owner, setting initial access policies and possessing full decryption rights.
- Consulting Doctors: Can request access for peer learning or second opinions. They receive a private key from the CA that may only allow "partial decryption"—for example, viewing clinical findings while masking the patient's personal identity.
- Patients: Can request their own records. The system generates a time-limited sharing link and a specific private key to ensure data portability without long-term exposure.
Security Analysis and Key Insights
The paper provides a qualitative analysis of the system's robustness:
- Multi-Barrier Defense: Access requires passing three hurdles: ID validation, a blacklist check, and possessing a private key that satisfies the ABE access tree.
- Resistance to Interception: Local encryption ensures data remains secure during transit and at rest, even if the cloud provider is compromised.
- Dynamic Revocation: The blacklist addresses the traditional difficulty of revoking ABE keys, providing a layer of real-time security.
Limitations and Future Work
While the framework is logically sound, several challenges remain:
- Computational Overhead: ABE is more computationally intensive than symmetric encryption. The impact on hospital latency when processing 60 GB of daily data requires further study.
- Centralized Trust: The system's security depends heavily on the integrity of the Certificate Authority.
- Implementation Specifics: Future research should focus on performance metrics, such as decryption time in emergency settings, and refining attribute granularity across diverse medical fields.
Summary updated