Educational illustration

Infrastructure walkthrough

Transport, application trust and operations solve different parts of the same problem. Step through the path or play the short walkthrough.

Why this matters. Changes fail at the dependencies nobody drew. Following one request shows where trust services and recovery owners sit.

Key terms 4

TLS: Transport Layer Security: the protocol that authenticates servers and encrypts traffic for HTTPS and many other services.
PKI: Public key infrastructure: the policies, systems and roles that issue, manage and revoke digital certificates.
OCSP stapling: A TLS server attaches a recent, CA-signed revocation response to the handshake, so clients do not have to ask the CA themselves.
Certificate Transparency: Public, append-only logs of issued certificates, specified in RFC 6962 and RFC 9162, that let domain owners see every certificate issued for their names.

An architecture illustration

The walkthrough explains dependencies. It does not represent deployed sites, capacity, routes or measured latency.

Step 1 of 5

Client request

A client begins a request to an application. Its identity, destination and protocol requirements establish the starting context. Name resolution comes first, so a resolver failure or a forged answer stops the request before any security control sees it.

Choose a stage or use the step controls to explore at your own pace.

Read the complete walkthrough

Client request
A client begins a request to an application. Its identity, destination and protocol requirements establish the starting context. Name resolution comes first, so a resolver failure or a forged answer stops the request before any security control sees it.
Routing & transport
Routing selects a path. Transport and TLS establish the connection; path selection alone does not establish application trust. The handshake depends on certificate validation, revocation checks and the trust anchors the client holds, so those services sit inline in every new connection.
Security policy
The agreed security controls evaluate the request. Policy ownership and failure behaviour should be explicit. A control that fails open keeps the service running without its protection; one that fails closed does the reverse. Either can be right if the choice is deliberate.
Application service
The application handles the request using its own authorisation and dependency controls. Its calls to databases, identity providers and third-party APIs are dependencies too, and each needs an owner and a failure plan.
Measurement & operations
Operators observe the agreed health signals, investigate failures and maintain the recovery process. Coverage determines what those signals can establish. Signals measured only from inside the network can miss what external clients experience, so where measurements are taken from is part of the design.