Business Continuity Plan & Incident Response
1. Recovery Objectives (RTO & RPO)
To ensure system reliability, Safe defines clear recovery targets for technical or infrastructure failures. These indicators allow the client's IT team to plan their operational tolerance.
| Service | Disruption Scenario | RPO (Recovery Point Objective) | RTO (Recovery Time Objective) | Priority |
|---|---|---|---|---|
| Database (SQL) | Instance or zone failure | 1 hour | 5 minutes | High |
| AI Engine | Lock-up or high load | Immediate | 30 minutes | Medium |
| Storage | Local hardware failure | 5 minutes | 5 minutes | High |
| Web App | Deployment error | N/A | 5 minutes | High |
Note: In Full On-Premise deployments, these metrics depend on the client's local hardware and backup infrastructure.
2. Incident and Security Breach Management
Safe has an Incident Response Team (IRT) led by its CTO and product lead, responsible for executing emergency protocols.
2.1. Severity Classification
• Level 1 (Critical):
Negative impact on reputation, client reaction or sensitive data leak.
• Level 2 (Major):
Significant operational disruption; data leak limited to public information.
• Level 3 (Limited):
Minor disruption with no impact on data security.
2.2. Response Protocol
• Identification:
Detection of abnormal behaviour, system errors or unauthorised access attempts.
• Containment:
Logical or physical isolation of affected resources to prevent the incident from spreading.
• Investigation:
Forensic analysis using snapshot backups to determine the root cause
• Recovery:
Restoring systems to their secure state and updating security controls.
3. Backup and Disaster Recovery (DR) Strategy
Data integrity is protected through a multi-zone redundancy strategy.
• Cloud (SaaS/Hybrid):
Databases and media files have synchronous replication to persistent disks.
• On-Premises:
The system persists logs and critical configuration data locally. Integration with the client's backup system or the use of hardware redundancy via mirrored NVMe disks is recommended.
• DR Testing:
An annual disaster simulation test is carried out to validate the adequacy of procedures and recovery KPIs.
4. Deployment and Maintenance Models
Safe offers total flexibility in how the system is installed and maintained, allowing the client to choose their desired level of control.
| Feature | Remote Configuration | On-Site Installation |
|---|---|---|
| Installation | Carried out via SSH or secure VPN by the Safe team. | Performed at the client's site with phone support. |
| Updates | Security patches and major versions managed by Safe. | Scripts and packages provided for the client to run. On-site visit for major updates (optional). |
| Maintenance | Periodic monitoring and incident handling included (optional). | Technical support available; servers maintained by the client. |
| Advantage | Lower operational load for the client's IT team. | Full control and absolute sovereignty over physical access. |
Regardless of the model, Safe guarantees that major version updates are tested in staging environments before moving to production.
Contact & Support
For technical questions about hardware compatibility or custom quotes, contact us at:
• Technical Support: support@safe.ai
• Sales: commercial@safe.ai