Transport Layer Security (TLS) strengthens the trustworthiness of application communications by ensuring that data exchanged over an untrusted network is not silently modified and is coming from the expected endpoint. While TCP provides delivery features such as sequencing and retransmission, TLS contributes to what many cybersecurity documents describe as “reliable” secure communication by adding cryptographic integrity protections . TLS uses integrity checks (such as message authentication codes in older versions/cipher suites, or authenticated encryption modes like AES-GCM and ChaCha20-Poly1305 in modern TLS) so that any alteration of data in transit is detected. If an attacker intercepts traffic and tries to change commands, session data, or application content, the integrity verification fails and the connection is typically terminated, preventing corrupted or manipulated messages from being accepted as valid.

This is distinct from merely being “stateful” (a transport-layer property) or “using TCP/IP” (a networking stack choice). TLS can run over TCP and relies on TCP for delivery reliability, but TLS itself is focused on confidentiality, integrity, and endpoint authentication . Public/private keys and certificates are used during the TLS handshake to authenticate servers (and optionally clients) and to establish shared session keys, but the ongoing protection that prevents undetected tampering is the integrity check on each protected record. Therefore, the best match to how TLS ensures secure, dependable communication is the message integrity mechanism described in option B.

When a system processes multiple information types with different security categorizations, cybersecurity standards require the system’s overall security categorization to reflect the highest impact level among those information types. This is commonly called the high-water mark approach. The reason is straightforward: the system is only as secure as the protection applied to the most sensitive or most mission-critical data it handles. If the system were categorized at the lowest impact value, an attacker could target the weaker control baseline and still reach higher-impact information, creating an unacceptable gap in confidentiality, integrity, or availability protection.

In practice, categorization evaluates the potential impact of loss for each of the three security objectives and then selects the highest level for each objective across all information types handled by the system. That resulting system categorization then drives control selection, assurance activities, and the rigor of monitoring and incident response expectations. This approach also supports consistent governance: it prevents under-protecting systems that contain a mix of low and high sensitivity information and aligns control strength with worst-case business impact.

Segregating data across systems can be a valid architecture decision to reduce cost or scope, but it is not the required categorization rule; it is an optional design strategy that must be justified and implemented securely. Merging categories or using the lowest value contradicts risk-based protection principles and would likely fail compliance and audit scrutiny.