Questions Regarding Flow-Based Antivirus Processing and Architecture
Hi ,
I would like to ask for some clarification regarding Flow-Based Antivirus behavior and architecture on FortiGate.
I reviewed the Administration Guide and several technical documents, but I could not find detailed explanations for some internal Flow AV processing behaviors.
I would highly appreciate it if you could clarify the following points or provide any related technical documentation, KB articles, technical tips, or architecture references.
- In Flow-Based Antivirus mode, is there still a file size limit / oversize handling mechanism similar to Proxy-Based AV, or is the oversize behavior only applicable to Proxy mode?
- In Flow mode, when scanning large files, does FortiGate bypass scanning after reaching a specific internal threshold, or can it continue scanning regardless of file size?
- Does Flow-Based Antivirus use the same Antivirus databases (Normal / Extended / Extreme) as Proxy-Based AV, or does Flow mode use a different AV engine/database architecture?
- If Flow mode does not fully buffer files like Proxy mode, how does the AV engine process files internally?
- Is scanning performed chunk-by-chunk?
- Does it use partial buffering/reassembly?
- Is detection primarily signature-based, IPS-engine-assisted, heuristic-based, or stream-pattern matching?
- Are there any technical references that explain:
- Flow AV internal architecture
- IPS engine interaction with Flow AV
- stream-based scanning logic
- buffer handling/reassembly mechanisms
- limitations and design differences between Flow and Proxy AV
I am trying to understand the internal processing logic and architecture differences between Flow-Based and Proxy-Based Antivirus in greater technical depth.
Any detailed explanation or internal/public technical references would be greatly appreciated.