1.3 Overview
The Remote Desktop Protocol: Multitransport Extension enables multiple side-band channels (also referred to as "multitransport connections") between an RDP client and server over different underlying transport protocols such as reliable UDP, or lossy UDP ([MS-RDPEUDP] section 1.3.1). Each multitransport connection leverages the strengths of the underlying transport protocol to efficiently deliver different types of RDP content, thereby improving the user's experience, especially on WAN or wireless networks.
After the main RDP connection has been established and secured, the server can initiate multitransport connections if it is determined that the connection would benefit from additional transports. Each multitransport connection that is initiated is bootstrapped with data that is exchanged on the main RDP connection by using the server-to-client Initiate Multitransport Request PDU ([MS-RDPBCGR] section 2.2.15.1) sent during the RDP connection sequence ([MS-RDPBCGR] section 1.3.1.1).
The Initiate Multitransport Request PDU contains information that uniquely identifies the multitransport connection; it contains a request ID, a cookie, and a protocol identifier that identifies the type of multitransport connection that the client attempts to establish. When the client receives the Initiate Multitransport Request PDU, it attempts to establish a secure multitransport connection with the server.
All multitransport connections are secured by using either Transport Layer Security (TLS) ([RFC2246], [RFC4346] and [RFC5246]) or Datagram Transport Layer Security (DTLS) ([RFC4347]). TLS is used to secure transport connections that ensure the reliable delivery of data, while DTLS is used to secure transport connections that can potentially lose data. If the creation of the underlying transport connection is successful and the TLS or DTLS handshake succeeds, then the multitransport connection is used to transport selected dynamic virtual channel traffic.