What the study found
The study found that an AI-driven hybrid architecture for multi-cloud storage can use telemetry-guided fragmentation to make data reconstruction harder. It also found that XGBoost and Random Forest had the highest predictive accuracy in the evaluations described.
Why the authors say this matters
The authors conclude that telemetry-driven adaptive fragmentation enhances predictive reliability and supports a resilient, zero-trust framework for secure multi-cloud storage. Here, zero-trust means the system does not assume cloud storage alone is enough to rebuild the data.
What the researchers tested
The researchers tested an AI-driven hybrid architecture that predicts fragment sizes from real-time telemetry, including bandwidth, latency, memory availability, and disk input/output. They evaluated it first with synthetic telemetry and then with hybrid telemetry that combined real Microsoft system traces and Cisco network metrics.
What worked and what didn't
Across the evaluations, XGBoost and Random Forest achieved the highest predictive accuracy, while Neural Network and Linear Regression performed moderately. The security validation indicated that partial-access and cloud-only attack scenarios could not reconstruct the data without the local vault fragments and the encryption key.
What to keep in mind
The abstract describes validation with synthetic telemetry and hybrid telemetry, not a full deployment study. It also does not provide numerical performance values or detailed limitations in the available summary.
Key points
- The study presents an AI-driven hybrid architecture for secure, reconstruction-resistant multi-cloud storage.
- Fragment sizes were predicted from real-time telemetry rather than fixed-size fragmentation.
- Data were compressed, encrypted with AES-128, and dispersed across independent cloud providers.
- Two encrypted fragments were kept in a VeraCrypt-protected local vault to block cloud-only reconstruction.
- XGBoost and Random Forest had the highest predictive accuracy in the reported evaluations.
- Security checks indicated that partial-access and cloud-only attacks could not reconstruct the data without the vault fragments and encryption key.
Disclosure
- Research title:
- Telemetry-driven fragmentation improved secure multi-cloud reconstruction resistance
- Authors:
- Munir Ahmed, Jiann-Shiun Yuan
- Institutions:
- University of Central Florida, University of Central Florida
- Publication date:
- 2026-01-27
- DOI:
- 10.3390/fi18020070
- OpenAlex record:
- View
Get the weekly research newsletter
Stay current with scholarly research without reading academic papers — one filtered digest, every Friday.