Fortinet FortiSandbox VM
AI-Driven Threat Analysis with Dynamic Behavioral Validation

Overview:

Powered by a native, state-of-the-art AI engine, FortiSandbox delivers real-time defense against evasive, previously unseen, and AI-driven threats. As cyber threats become increasingly evasive and AI-driven, organizations need defenses that continuously adapt to new attack techniques. FortiSandbox combines advanced malware analysis, continuously evolving machine learning, and behavioral intelligence to detect malicious intent with high confidence and near-zero false positives. Most files are classified within seconds using AI-driven inspection, while deeper behavioral analysis is selectively applied to uncover sophisticated and evasive threats.

Supporting multiple operating systems and file types, FortiSandbox provides actionable threat intelligence and detailed reporting to accelerate investigation and response. Native integration with the Fortinet Security Fabric and a broad ecosystem of security tools enables coordinated threat prevention across email, endpoints, web traffic, network shares, and the network edge.

FortiSandbox VM (FSA-VMS) brings this same AI-driven detection engine to a virtual appliance, giving organizations the flexibility to deploy sandboxing on their own hypervisor infrastructure. Universal VM is an all-in-one license that lets you choose any local, cloud, or custom virtual machine type and operating system, detaching VM licenses from OS licenses to reduce licensing complexity.

Highlights:

What's New in FortiSandbox v5.2

Stronger Security. Smarter Operations. Greater Control. Delivering powerful new capabilities to help you secure, manage, and scale with confidence.

---

Enterprise Trust: Validated Security and Compliance

NIAP — FortiSandbox is listed on the NIAP PCL as product 11636, evidencing independent evaluation against rigorous U.S. government security assurance standards and reinforcing its suitability for government, defense, and critical infrastructure environments.

HIPAA-Compliant — FortiSandbox aligns with strict requirements for safeguarding protected health information (PHI), supporting secure handling, processing, and analysis of sensitive healthcare data.

SOC 2® Certified — FortiSandbox is SOC 2® certified, demonstrating independent validation of its security, availability, and confidentiality controls.

Features:

Advanced Threat Protection

• Advanced AI to identify zero-day threats faster and with better detection
• Inline blocking to detect and protect against zero-day malware, including ransomware; blocks and holds malicious content at the FortiGate and sends it to the sandbox for analysis/verdict
• Real-time identification of zero-day phishing sites, including spam and malware-hosted sites
• Network threat detection in sniffer mode — identifies botnet activities, network attacks, and malicious URL visits
• Threat enrichment through FortiGuard IOC
• Sandbox Community Cloud for shared analysis within the worldwide community of FortiSandbox deployments

System Integration Support

• Integrated mode with FortiGate: HTTP, SMTP, POP3, IMAP, MAPI, FTP, IM, and their equivalent SSL-encrypted versions
• Integrated mode with FortiMail: SMTP, POP3, IMAP
• Integrated mode with FortiClient EMS: HTTP, FTP, SMB
• Integrated mode with FortiWeb: HTTP
• Sniffer mode: HTTP, FTP, POP3, IMAP, SMTP, SMB
• Proxy inspection via ICAP
• MTA/BCC mode via SMTP
• NetShare Scan mode via FTP, sFTP, CIFS, NFS, OneDrive, AWS S3 Buckets, Azure Blob, and Google Cloud Storage
• Dynamic threat intelligence DB updates of malicious file checksums and URLs
• JSON API to automate uploading samples and downloading actionable malware indicators for remediation
• Remote and secured logging with FortiAnalyzer, FortiSIEM, CEF servers, and syslog servers

Deployment

• File submission from integrated device(s)
• Sniffer mode deployment with TCP RST support to reset a client's connection with the suspicious server
• Network Share Scan with large file support (e.g., ISO images, network shared folders, SMB/NFS, AWS S3, and Azure Blob)
• Proxy adapter submission with multi-tenancy support
• OT deployment with supported services: BACnet, HTTP, IPMI, Modbus, S7comm, SNMP, TFTP
• High availability with primary and secondary nodes for redundancy
• Port monitoring for cluster fail-over
• Clustering up to 99 worker nodes for higher throughput
• Air-gapped network support
• Aggregate interface support for increased bandwidth and redundancy
• Isolated administrative traffic from VM image traffic

Advanced AI Scan (Static AI Scan) Features

• Integrated with the new Advanced AI engine and model
• Integrated with the full FortiGuard Antivirus database of heuristic and checksum signatures
• Intelligent adaptive scan profile that optimizes sandbox resources based on submissions
• Parallel scan to run multiple distinct VM types simultaneously
• Extracts and scans files embedded in documents
• Extracts and scans URLs embedded in documents and QR codes
• Extracts and scans images in documents using OCR
• Integrates with third-party Yara rules
• Cloud query for the latest known malware and clean files
• File checksum allowlist and blocklist options
• Scans URLs from submitted emails and files
• Rating Engine Plus that leverages the latest FortiGuard ML rating
• VM scan ratio for efficient utilization of VMs

Sandboxing VM (Dynamic AI Scan) Support

• AI-powered behavioral analysis constantly learning new malware and ransomware techniques
• Concurrent dynamic analysis VM instances
• OS types supported: Windows 11/10/8.1/7, macOS, Linux, Android, and ICS systems
• Customizable VMs for Windows and Linux OS
• Configurable internet browser support, including Internet Explorer, Microsoft Edge, Google Chrome, and Mozilla Firefox
• Sandbox interactive mode with video-recording of malware interaction and VM screenshots
• Nested VMs for on-premise and cloud deployment
• Anti-evasion detection techniques: API obfuscation, bare-metal detection, command and control, direct system calls, execution delay, memory-only payload, process hollowing/injection, runtime encryption/packing, system fingerprinting, time bomb, user files check, user interaction check, and VM/sandbox detection
• Callback detection for malicious URL visits, botnet C&C communication, and attacker traffic from activated malware
• Downloadable captured packets, tracer logs, and screenshots
• File type support: Windows executables, Microsoft Office, document/email, Android files, Linux files, macOS, and web files
• File compression support
• User-defined extensions

Monitoring and Reporting

• AI-based threat summary using collected indicators and results
• Dashboard widgets for connectivity and services, license status, scan performance, and system resources
• Scan performance page for tracking historical usage
• Real-time monitoring widgets: scanning result statistics, scanning activities over time, top targeted hosts, top malware, top infectious URLs, and top callback domains
• Drilldown event viewer with a dynamic table of actions, malware name, rating, type, source, destination, detection time, and download path
• Reports and logging via GUI, downloadable PDF, and raw log file
• Detailed job report generation
• Periodic logs of system status, performance, scan statistics, and system resource usage
• MITRE ATT&CK v11 support
• Downloadable tracer logs, PCAP, and indicators in STIX 2.0 format
• Notification emails when a malicious file is detected
• Weekly reports to global email lists and administrators
• TAC report for a comprehensive snapshot of system configuration and status

Administration

• Configuration via GUI and CLI
• Multiple administrator accounts supporting full or view-only access
• RADIUS authentication for administrators
• Single sign-on via SAML
• Self-check widget for configurations, connectivity, and services
• Cluster management page for administering HA and cluster nodes
• Centralized search page allowing administrators to build customized search conditions
• Upload any license from a single convenient page
• VM status monitoring
• Automatic engine and signature updates
• Automatic check for new VM image availability
• System health check alerting
• NTP via FortiGuard support
• Backup, restore, and revision of system configuration
• Consolidated CLI for troubleshooting
• Option to auto-submit suspicious files to a cloud service for manual analysis and signature creation
• Option on NetShare scan mode to prioritize and forward files to a third party for further scanning

Specifications:

FortiSandbox VM is offered under the FSA-VMS subscription license, bundling the Advanced AI engine and scaling from FSA-VMS1 up to FSA-VMS4. Each tier is expandable to a maximum of 200 cloud VMs and is tested on-premises on a Hyper-V server with the corresponding CPUs, memory, and VMs.

	FSA-VMS1	FSA-VMS2	FSA-VMS3	FSA-VMS4
Capacity
Local VM Capacity	0–8	0–16	0–32	0–64
Cloud VM Expansion1	1–200
Performance and Capacity
Effective Sandboxing Throughput2 (Files/Hr)	8,0004	24,000	48,000	96,000
Static Analysis Throughput5 (Files/Hr)	20,0004	60,000	120,000	240,000
Dynamic Analysis Throughput6 (Files/Hr)	2004	400	800	–
FortiMail Throughput7 (Emails/Hr)	80,000	240,000	480,000	960,000
MTA Adapter Throughput (Emails/Hr)	20,000	60,000	120,000	240,000
Sniffer Mode Throughput (Gbps)	1	TBD	TBD	TBD
Number of Users8	1,000	2,000	4,000	8,0009
Platform
Hypervisor Support	Hyper-V
Universal VM Licensing	Yes — detaches VM licenses from OS licenses

1. The ranges reflect Universal VM support through firmware version 5.0 and later.
2. Tested based on files with 80% documents and 20% executables; measured based on v5.2. Includes both static and dynamic analysis with pre-filtering enabled.
4. FSA-VMS1 is tested on a Hyper-V host with 16 vCPUs, 32 GB RAM, and 8 VMs.
5. Includes receiving, job handling, AV engine, Yara engine, and Cloud Query; measured based on v5.2.
6. Tested with Static Analysis, and all files are forwarded to Dynamic Analysis.
7. Based on a ratio of one email with attachment to 10 emails.
8. Based on a ratio of one user per 80 emails on a 10-hour period with 10% on Dynamic Scan.
9. FSA-VMS4 throughput is constrained by the 96 vCPU limit in version 5.0.

Documentation:

Download the Fortinet FortiSandbox Series Datasheet (PDF).