SDP (Session Description Protocol) is the small text format that describes what a media stream contains, so two endpoints can agree on how to handle it. It is not a transport — it carries no video — it is the "menu" exchanged during stream setup: which codecs are in use (H.264, H.265), the payload types, resolution and timing hints, and the transport details. Standardised by the IETF (RFC 8866, updating the earlier RFC 4566), SDP is what an RTSP DESCRIBE response hands back to the client.
In a surveillance pipeline SDP is the quiet glue between RTSP and RTP. When the VMS asks a camera to describe its stream, the camera replies with SDP; the VMS reads it to learn the codec and payload format, then sets up RTP to receive the media accordingly. Without that description, the receiver would get a flow of packets it does not know how to interpret.
Practically, SDP is rarely something an integrator configures by hand, but it is where codec-mismatch problems surface. If a camera advertises a codec or payload type the VMS does not support, or describes the stream incorrectly, the symptom is a session that sets up but yields no decodable video. Diagnosing such "connects but no picture" faults often means inspecting the SDP. The deep mechanics belong to the Video Streaming section; in surveillance, SDP is simply how the stream announces what it is.
