The camera is added to the VMS via ONVIF and video comes through; but the PTZ doesn’t respond to the joystick, motion events don’t arrive, and the recording the camera wrote to its SD card during a network outage isn’t pulled back. Usually this isn’t a fault: although both devices are “ONVIF compliant”, they support different profiles or different optional functions of the same profile. ONVIF (Open Network Video Interface Forum) is an organization that develops interface standards so that IP-based physical security products from different manufacturers can work together. The standards are implemented as sets of functions called “profiles”. Below are the profiles most commonly encountered in video surveillance, recent changes and how to verify compatibility on site.
The profile concept: mandatory and optional functions
Each profile defines mandatory functions required to claim conformance, and conditional/optional functions that must be implemented according to the standard if supported. That’s why basic video streaming may work fine between two “ONVIF compliant” devices while functions such as PTZ, audio, events or analytics don’t. In corporate projects, compatibility is always verified by model, firmware and VMS version. A logo isn’t enough.
The distinction is two-sided. Profiles are declared separately for the device (camera, NVR) and the client (VMS, the side of the NVR facing the camera). A camera supporting Profile G doesn’t mean the VMS can pull back those recordings; the VMS must also have implemented this function as a Profile G client.
Profile S: basic video streaming and the end of support
For many years, Profile S was the most common profile for IP cameras: video and audio streaming, PTZ control, basic configuration. On 9 October 2025, ONVIF announced it would deprecate Profile S and recommended Profile T as its successor. According to the organization’s Q&A page, after 31 March 2027 new products or new software/firmware versions can’t be submitted for Profile S conformance. The reason is that the username token authentication mandated by Profile S is too weak for today’s cybersecurity expectations.
Existing compliant devices don’t automatically stop working on that date; their conformance lasts until the manufacturer withdraws its declaration. For new purchases, the required Profile T, G or M functions are verified on both the device and client side. Practical advice: write Profile T into the specification for new purchases, use digest authentication or TLS/HTTPS on existing devices where possible, and add devices that no longer receive security updates to the lifecycle plan.
Profile T: advanced video streaming
Profile T covers almost all Profile S functions and extends them for current needs. According to ONVIF, the highlights are:
- H.264 and H.265 video encoding support
- Digest authentication; media transfer over HTTPS where supported
- Configuration of imaging settings
- Motion alarm and tamper events
- Metadata streaming; audio output and PTZ on supported devices
Profile T doesn’t include username token authentication, IP address filtering or older formats such as MJPEG and MPEG-4. If an old VMS version works only with Profile S logic, it may not see the H.265 stream on a new Profile T camera, and you’ll have to fall back to an H.264 profile. The VMS version is checked before buying cameras.
Profile G: recording, search and playback
Profile G covers a client searching and playing back recordings on recording devices (SD card on the camera, NVR) and managing recording configuration. In practice it matters in two scenarios: the VMS pulling back the edge recording the camera made to its own storage during a network or server outage to fill the gap, and accessing recordings on different-brand recorders through a common interface. This feature must also be supported on the VMS side.
The edge recording scenario shouldn’t be trusted without testing on site. Unplug the camera’s network cable for a few minutes, plug it back in and check whether the VMS fills that gap from the SD card. The card’s capacity, wear endurance and the accuracy of the camera clock are also part of this scenario; a recording pulled back from a camera whose clock has drifted is written to the wrong time slot.
Profile M: analytics metadata and events
Profile M defines the standard transfer of metadata produced by cameras or analytics devices (object class, position, counts, plates and so on) and analytics events, and the configuration of analytics rules. Bringing a “person entered the zone” event produced by the camera’s AI into a different-brand VMS, or using it in smart search, may be possible thanks to this profile; however, supported metadata types differ by product. Profile conformance doesn’t mean the target VMS can process all analytics events.
A common situation on site is that the event reaches the VMS but the class information is lost: the VMS shows the event as a generic “motion”, and the “vehicle” or “person” distinction isn’t visible. This shows the integration is incomplete, and it’s only noticed if checked per event type in a test matrix.
Other related profiles
| Profile | Scope | Use in video surveillance |
|---|---|---|
| T | Advanced video streaming | The basic choice for new IP cameras |
| G | Recording, search, playback | Edge recording, NVR integration |
| M | Analytics metadata and events | Bringing AI events into the VMS |
| A / C | Access control configuration / monitoring | Access control integration |
| D | Access control peripherals | Peripherals such as readers and locks |
The difference between RTSP and ONVIF
RTSP is a video streaming protocol; it lets you get video from a camera. ONVIF defines management functions such as device discovery, configuration, selection of stream profiles, PTZ, events and metadata, and also uses RTSP/RTP for streaming. Adding a camera to the VMS with only its RTSP address provides video but doesn’t bring event, PTZ and configuration functions. Having video doesn’t mean the integration is complete.
Common ONVIF connection errors
- Clock difference: ONVIF authentication uses timestamps. If the clock difference between the camera and the VMS is large, the connection may be rejected despite the correct password. Check the NTP settings on both sides first.
- ONVIF service or user disabled: With some manufacturers, ONVIF is disabled by default or requires an ONVIF user separate from the web interface user.
- Discovery doesn’t cross VLANs: ONVIF device discovery (WS-Discovery) uses multicast and doesn’t reach a different subnet behind a router. If the camera network is in a separate VLAN, devices are added manually by IP address.
- Wrong stream profile: The VMS takes the camera’s first media profile; if that profile is the substream, recording is done at low resolution. After the camera is added, check which profile is used for recording.
How do you verify compatibility?
- Check the profiles the model and firmware version declare in the ONVIF conformant products database.
- Look at the VMS manufacturer’s supported device list and ONVIF driver notes.
- For critical models, try streaming, PTZ, events, edge recording and metadata functions one by one in a test environment and record the results in a matrix.
- Repeat the test after firmware or VMS updates.
For the profile list and the official product database, ONVIF’s own website (onvif.org) is the primary source; the ONVIF mention in a manufacturer brochure shouldn’t be accepted without comparing it with this database.
Frequently asked questions
Will my Profile S devices keep working?
Yes. According to ONVIF, the end of Profile S doesn’t affect the operation of Profile S systems in the field; however, after 31 March 2027, new products or new software/firmware versions can’t be submitted for Profile S conformance, and dropping the old username token authentication is recommended.
What does “ONVIF compliant” guarantee?
That a product supports the mandatory functions of the profile it declares. Optional features, manufacturer-specific analytics and behavior with specific VMS versions must be tested separately. Conformance should be verified in the official ONVIF product database.
Can I bring the camera’s AI events into any VMS with Profile M?
Both the camera and the VMS must support Profile M and process the relevant metadata types. The scope of support varies by product; integration testing is essential.
In mixed-brand systems, we do this verification with a per-model compatibility matrix as part of our ONVIF and RTSP integration service.