Transport encryption
Protects a channel between two transport endpoints, typically client and server. TLS 1.3 provides confidentiality, integrity and server authentication for that channel.
Between which endpoints is plaintext excluded?
DIFFERENT PROTECTION OR OPERATING MODEL
The mechanisms are not alternatives at the same layer. Sound E2EE systems also use secure transport because E2EE does not cover every header, handshake artefact or availability attack.
Evaluation starts at actual protection and operating endpoints, not marketing terminology.
Protects a channel between two transport endpoints, typically client and server. TLS 1.3 provides confidentiality, integrity and server authentication for that channel.
Protects messages at the application layer so intermediary servers should not decrypt content; keys remain at communicating endpoints.
Each row names the practical difference and why it matters during evaluation.
| Criterion | Transport encryption | End-to-end encryption | Why it matters |
|---|---|---|---|
| Protection endpoints | Per transport connection, e.g. app ↔ server | Communicating end devices | The word encrypted is insufficient without naming endpoints. |
| Server access | Server is a transport endpoint and can process application data | Server should only relay ciphertext | Features such as server-side search need a separate model. |
| Key responsibility | Certificates and session keys at the service endpoint | Identity and message keys at clients | Device security becomes part of content confidentiality. |
| Intermediate storage | Service may store plaintext after TLS termination | Service generally stores encrypted content | Backups and indexes must match the same protection model. |
| Multi-device operation | A new login can be sufficient for server access | Each device needs identity, key state and delivery | Adding a device is a security event. |
| Metadata | Transport does not hide all network and service metadata | E2EE protects content, not automatically routing and usage patterns | Metadata protection needs separate controls. |
HTTPS/TLS commonly terminates at the server. Whether the server can read message content is decided at the application layer.
RFC 9420 still recommends secure transport because observation and manipulation of delivery can expose information or enable availability attacks.
These product-neutral questions can be used directly in procurement or architecture review.
Where exactly does each encryption layer terminate?
Can server code or administration decrypt message content?
How are new devices authenticated and keys distributed?
What happens on device loss, recovery or key revocation?
Are export, search, preview and backup part of the same protection model?
No. TLS 1.3 strongly protects a channel. It simply answers a different trust question from application-level end-to-end encryption.
In robust systems, yes: transport protects additional protocol data, makes manipulation harder and reduces observability that content encryption alone does not fully address.
VENTEX links lead to current product state and known boundaries. External links lead only to the primary sources used here.