CompTIA SY0-701 Bundle

One unsalted SHA-256 operation is deliberately fast and deterministic, which would let attackers use GPUs to evaluate enormous candidate lists quickly. Removing the existing salts also reintroduces equal-hash leakage for equal passwords. SHA-256 is a useful general-purpose digest, but password storage requires a suitable intentionally expensive password-hashing or KDF scheme, not a single inexpensive digest computation.

Failed-login lockouts limit online attempts made through an application's authentication endpoint and can help against remote guessing. The attacker described here does not submit login requests; it already has salted verifier hashes and computes candidate matches locally. No server-side failed-login counter is incremented during those offline guesses, so increasing lockout severity will not materially raise the cost of each guessed password.

TLS encrypts a network connection and protects data in transit between an application and database service when correctly configured. That reduces interception risks but does not affect the cost of cracking verification hashes already stolen from storage. The question states the database has been exfiltrated, so changing transport protection while keeping a fast password hash leaves the attacker’s offline GPU advantage intact.

An HSM can protect cryptographic material against extraction, but a full recovery key controlled by one administrator would still allow that custodian to authorize or initiate recovery alone, depending on the access design. This option does not impose the board's required independent contributions. Hardware protection and multi-person reconstruction solve different problems; a properly configured HSM could support a threshold arrangement, but this single-owner design does not.

Retaining a complete recovery-key copy in escrow with one custodian permits that person to recover the key before another person knows. Sending an alert afterward is detective oversight, not a cryptographic requirement for two independent contributors. The board wants prevention of single-person reconstruction, which calls for split knowledge or enforced multi-party key release rather than retrospective notification.

A written change approval creates an administrative record that a proposed recovery action was authorized. If one custodian continues holding and using the entire key, that individual retains the technical ability to reconstruct it without another party's participation. The policy explicitly prohibits such unilateral capability, so a paperwork condition without distributed key material or strong joint control is insufficient.

A software-encrypted private-key file on an application server can protect the key at rest, but if the same service account holds the decryption password, a compromised application process may decrypt or extract the signing key into ordinary operating-system memory. The question explicitly prohibits that exposure and requires signing within an isolated cryptographic boundary. Local file encryption with colocated credentials does not meet that standard.

A Trusted Platform Module on an employee's endpoint binds cryptographic operations to that particular local device, often protecting disk-unlock or device-attestation keys. Distributing the service-wide financial signing key to employee laptops creates extra endpoints and does not centralize signing inside the protected service boundary. TPMs can support device trust, but they are not the specified architecture for centrally controlled, nonexportable transaction signing.

Splitting key shares among administrators can limit who can reconstruct the secret, which may improve organizational custody of recovery material. However, this option explicitly recombines the entire signing key in normal application process memory for each signature. That violates the nonexportability and hardware-boundary requirement once reconstruction occurs, even if two custodians were involved beforehand.

A network-attached or rack-mounted HSM can protect keys for server applications and perform remote cryptographic operations under managed access. It is appropriate for centralized signing services, but the question requires trust material tied to a particular laptop and its measured boot state. A remote service key is not inherently sealed to that endpoint's firmware and boot measurements.

A password-salt database stores random per-account salt values used when hashing passwords so that equal passwords do not yield equal verifier records. Those values are not hardware trust anchors and do not measure firmware or boot state. A shared authentication-service database cannot release a laptop disk key specifically in response to that device's local measured configuration.

Ephemeral TLS key exchange such as ECDHE establishes short-lived traffic secrets for a network session and can provide forward secrecy if long-term server signing keys are later exposed. It does not persistently bind a disk-unlock secret to a laptop's boot measurements or supply the local hardware root of trust described. Network-session confidentiality and endpoint disk-key sealing solve separate problems.

Short-lived certificates reduce the duration for which a compromised server identity credential is accepted, and regular certificate replacement can help manage exposure. However, if recorded sessions used static RSA key transport without forward secrecy, later theft of the relevant long-term private key can reveal those captured session secrets. Certificate expiration alone does not change how past traffic keys were established.

An online certificate status responder helps a client determine whether a server certificate has been revoked, supporting trust decisions during a handshake. Revocation status does not change the mathematical derivation of historical TLS traffic keys and cannot retrospectively provide forward secrecy. Even perfect OCSP checking would not protect recorded static-key-transport sessions against later exposure of the private decryption key.

Public-key pinning restricts which server keys a controlled client will accept, reducing the risk of connecting to an unauthorized certificate or key even if a CA issues a valid-looking chain. It does not make the TLS session secrets ephemeral. A pinned server with static RSA key transport could still expose recorded sessions after later long-term key theft, so key agreement mode is decisive.

Trusting any certificate that chains to a publicly trusted CA follows the browser-style public PKI model, which checks chain, identity and validity. The company's threat model requires a stricter local key allowlist, so ordinary CA trust could still accept another valid certificate signed for that service using a different public key. It would not enforce the particular expected server key material.

Ephemeral ECDHE protects session confidentiality through fresh key agreement and may provide forward secrecy against later theft of the server's signing key. It does not identify which server public keys the native application is allowed to trust. An attacker presenting a different CA-valid key could still satisfy a normal TLS handshake unless separate identity/pinning checks reject the key.

OCSP indicates certificate revocation status for a queried certificate and may be part of ordinary PKI validation. It does not compare the server's public key to an independently managed application allowlist. A CA-valid, not-revoked certificate with an unexpected key could still be accepted unless a pinning policy explicitly rejects it. Revocation checking and identity pinning address different trust questions.

Public-key pinning requires a managed client to recognize a particular server key or allowed key set; it can reject otherwise CA-valid certificates with unexpected keys. That mechanism does not return the issuer's current revocation state for an arbitrary certificate. It solves server-key identity enforcement for a controlled application, whereas the question specifically asks about online certificate revocation without a full list.

A certificate revocation list is a signed issuer-published set of certificates revoked before expiration, which a validator may download and inspect. It is useful for disconnected or batch-oriented validation when lists can be periodically distributed. However, the stated client wants a status result without retrieving the complete list, making targeted OCSP checking more directly aligned with the network and data-transfer requirement.

A certificate signing request carries identifying information and a subject public key to a certificate authority when asking for certificate issuance. It is not a live certificate-status query and does not tell a relying party whether an already issued certificate has been revoked. CSR processing happens earlier in the certificate lifecycle, before the certificate is used in trust decisions.

OCSP is designed for online queries about individual certificate status and usually expects access to an authorized responder or stapled response. The appliance is disconnected and receives signed lists through a scheduled delivery process, so it cannot depend on obtaining fresh interactive OCSP results. OCSP can be effective for connected clients, but it does not match this offline distribution design.

Ephemeral ECDHE is a TLS key-agreement mechanism that can give past sessions forward secrecy against later compromise of a long-term key. It does not contain a list of revoked certificate serial numbers or inform a disconnected validation appliance which issuer credentials have been invalidated. The question is about certificate trust-status data rather than transport session secrecy.

A trusted root and a valid chain establish that the certificate can be related cryptographically to a recognized issuing authority. Chain construction alone does not establish whether a certificate was revoked after issue. A disconnected validator still needs suitable revocation data, such as the periodic signed lists described, to make that additional status decision.

A certificate revocation list communicates which already-issued certificates an issuer has revoked. It is useful to a relying party deciding whether a deployed certificate should still be trusted, not to a server that wants a new certificate issued. The service owner has a freshly generated key pair and needs to submit an authenticated certificate request before any issuer-signed certificate exists.

OCSP permits online checking of an issued certificate's revocation status against an authorized responder. It is part of validation after a certificate has been issued and used. It does not convey a new key pair's subject information or provide the proof of possession and attributes required by an issuing certificate authority for a new server certificate.

A trusted root and a valid certificate chain are used by clients to verify that an already-issued server certificate descends from an accepted certificate authority. They do not themselves generate or submit a request for issuance. Once the CA has issued the service certificate, its chain will matter to browsers and applications, but the immediate lifecycle artifact is the CSR.

A certificate signing request is used when a server owner asks a CA to issue a certificate for a public key. It cannot by itself make a browser trust an issued certificate from an unknown authority. The described certificate already exists and matches the subject, so the validation gap lies in the trust chain or anchor rather than in generating another request artifact.

Certificate pinning can add a stricter client-specific key expectation beyond ordinary CA trust, but it does not fix generic browser trust for a server certificate with an unknown issuing authority. Standard browsers will reject an untrusted chain before treating pinning as a substitute for authorization. Pinning is useful only where a controlled client can manage its explicit allowed key set safely; the stem asks for basic issuer trust.

OCSP checks whether an issued certificate has been revoked by its CA. It cannot make an unknown issuing root trustworthy or repair a missing intermediate certificate. A certificate can be unrevoked and still lack a chain to a browser's trust anchor. The appropriate first task is to establish trusted issuance and a valid chain, then apply revocation and other validation checks.

A private digital-signing key represents the individual's signing identity and can be used to generate signatures attributable to that identity. Retaining it for later impersonation is not the same as recovering encrypted corporate archives and undermines the separation between data recovery and signature authenticity. If signing credentials are compromised or no longer needed, policy may require revocation, but they should not become enterprise recovery keys for unrelated encryption.

Certificate revocation communicates that a certificate should no longer be accepted for certain trust decisions. It does not recreate a lost symmetric encryption key, derive plaintext from ciphertext or supply the recovery material needed to decrypt archived records. Revocation is a lifecycle and trust-control process, whereas the question calls for preserving authorized access to older encrypted data when an employee is absent.

Rotating current encryption keys protects new operations or limits future exposure, but old archives encrypted with a previous key remain unreadable unless that historical key or a supported rewrapping path is retained. The problem specifically concerns already-encrypted corporate records after the employee leaves. Rotation without planned recovery or key retention cannot decrypt historical ciphertext and would risk permanent data loss.