End-to-end security

Description

End-to-end security is a 3GPP security concept defined for IMS-based media services, referring to media protection that extends between two IMS UEs without being terminated by any intermediary. This means that the security association is established directly between the two endpoints, and no intermediate node—including IMS core nodes, media gateways, or application servers—can decrypt or modify the media. The protection covers confidentiality and integrity, and it is designed to provide the strongest possible privacy for user communications. The concept is specified in 3GPP TS 33.328 (IMS media plane security) and related specifications such as TS 23.333 (MRF), TS 23.334 (IMS-AGW), and TS 29.162 (interworking).

In an end-to-end security setup, the two UEs negotiate security parameters and keys during session establishment, often using protocols like MIKEY (Multimedia Internet KEYing) or SDP security descriptions. The media is then encrypted and integrity-protected using SRTP (Secure Real-time Transport Protocol) or similar mechanisms. Because the protection is end-to-end, any intermediate node that needs to process the media (e.g., for transcoding, conferencing, or lawful interception) cannot do so without breaking the security. Therefore, end-to-end security is typically used when such processing is not required, or when the endpoints are capable of handling the security themselves.

The architecture involves the UEs, the IMS core network (P-CSCF, S-CSCF, etc.), and the media path. The signalling to establish the security association is carried over the IMS SIP signalling plane, which may itself be protected hop-by-hop. The media plane security is then applied end-to-end. The UEs must support the necessary security protocols and algorithms, and the network must allow the negotiation to proceed. In some cases, the network may need to be aware of the security parameters for policy control or charging, but it does not have access to the media content.

Key components include the UEs, the key management functions, and the security protocols. The UEs are responsible for generating and managing the keys, and for applying encryption and integrity protection to the media packets. The network provides the signalling path and may assist in key agreement if required, but it does not terminate the security. The specifications define the procedures for negotiating security capabilities, exchanging keys, and handling errors. They also specify how end-to-end security interacts with other security mechanisms, such as end-to-access edge security and hop-by-hop security.

End-to-end security is important for applications that require the highest level of privacy, such as confidential business calls or personal communications. It prevents any intermediary, including the network operator, from accessing the media content. However, it may conflict with lawful interception requirements, so its use may be restricted in some jurisdictions. The concept has been stable across releases from Rel-9 to Rel-19, with no significant changes to the definition, indicating a mature and well-understood security mechanism.

Purpose & Motivation

End-to-end security was created to provide the highest level of privacy for IMS media communications by ensuring that only the communicating parties can access the media. Before its introduction, media could be protected only on a hop-by-hop basis or between the UE and the network edge, which meant that intermediate nodes could potentially access the content. For users who require absolute confidentiality, such as government agencies or enterprises handling sensitive information, end-to-end security is essential. It addresses the limitations of other security models by eliminating any trusted intermediary in the media path.

The motivation also stems from the need to protect against insider threats and to comply with privacy regulations that mandate end-to-end encryption for certain types of communication. While end-to-access edge security protects the access leg, it still allows the core network to access the media for lawful interception or transcoding. End-to-end security removes that capability, which is a trade-off between privacy and network functionality. The 3GPP specifications provide a standardized way to implement end-to-end security, ensuring interoperability between different vendors and operators.

Historically, the concept was introduced in 3GPP Release 9 as part of the IMS media plane security enhancements. It built upon earlier work on IMS security and was designed to complement other security layers. The specifications TS 33.328 and TS 23.334 define the architecture and procedures. Since then, the definition has remained unchanged, indicating that the fundamental approach has been stable and continues to be relevant in modern networks including 5G.