Architecture and Deployment
Resiliate™ is a modular, AI-native filesystem designed for autonomous data defense. Its architecture integrates kernel-level modules, AI inference engines, and control interfaces to provide real-time data integrity and threat mitigation.
Core Components
- AiFS: A FUSE-based filesystem that enables seamless integration with existing systems while providing advanced features like self-healing and anomaly detection.
- CellFS Kernel Module: A DKMS module that operates at the kernel level to monitor and manage low-level file operations, ensuring data integrity and security.
- Cybernetic Engram (CE) Engine: An AI-driven engine that evaluates the state of data, users, and processes to make informed decisions about file operations. Learn more at docs.saf.ai/ref/ce.html.
- Control Plane: Provides a unified interface for interacting with the CE engine, plugins, back-end stores, and statistical models. Detailed information is available at docs.saf.ai/ref/control-plane/index.html.
Deployment Options
- Linux Integration: Resiliate™ can be deployed on Linux systems, integrating with existing filesystems and providing advanced data protection features.
- Windows Integration: Supports integration with Windows environments, including Active Directory. Refer to docs.saf.ai/guide/windows/active-directory.html for setup instructions.
- CEPH Storage: Experimental support for CEPH storage clusters allows for scalable and distributed storage solutions. Setup guidance is provided at docs.saf.ai/guide/ceph/cluster.html.
Installation and Configuration
To install Resiliate™:
- Add the sāf.ai repository to your system's package manager.
- Install the Resiliate™ package using your package manager.
- Configure the system using the
aifs
command-line tool. Detailed instructions are available at docs.saf.ai/guide/install/index.html.
For configuration options and commands, refer to the CLI documentation at docs.saf.ai/cli/index.html.
Testing and Validation
After installation, it's crucial to test the system to ensure it's functioning correctly:
- Data Creation and Deletion: Create and delete files to observe Resiliate™'s behavior.
- Data Recovery: Test the self-healing capabilities by simulating data corruption.
- Analysis: Use provided tools to analyze the system's responses and logs.
Testing procedures are detailed at docs.saf.ai/guide/test/.
Deployment Options
Resiliate™ is designed to run flexibly across a variety of environments, from tightly controlled air-gapped datacenters to dynamic, containerized cloud-native stacks.
On-Premises Deployments
- Linux Servers: Full kernel-level integration via CellFS for mission-critical workloads.
- Virtual Machines: Supports KVM, VMware, and Hyper-V environments with persistent CE states.
- Air-Gapped Networks: Resiliate™ can operate fully offline with pre-seeded engrams and signed models.
- Bare Metal Configurations: Optimized for environments requiring high-throughput, low-latency forensic pipelines.
Learn more about on-premise deployment architecture:
docs.saf.ai/guide/on-prem/index.html
Cloud-Based Deployments
- Kubernetes / K3s Clusters: Native Helm charts and CRDs allow you to run Resiliate® as a sidecar or CSI plugin.
Deployment guide: docs.saf.ai/guide/kubernetes/ - AWS, GCP, Azure: Bring-your-own-storage model enables use with EBS, S3, GCS, Blob. Engram states remain decoupled from physical cloud volume identity.
- SaaS-like Control Plane: Optional hosted control plane for observability and coordination, without compromising on local autonomy.
Full cloud deployment documentation:
docs.saf.ai/guide/cloud-deploy/index.html