Skip to content

SAP-C03 · Domain 2: Security, Compliance, and Governance · 25% of the exam

Task 2.2: Design solutions to implement encryption and key management strategies.

Encryption at rest with customer managed KMS keys and rotation, encryption in transit with ACM and AWS Private CA, dedicated HSMs where regulation demands them, and planning for post-quantum cryptography.

Study it

  • Keys and certificates: KMS key policies, rotation, multi-Region keys, ACM and AWS Private CA

    Partly covered by: KMS, Envelope Encryption & Private CA

  • Dedicated HSMs and post-quantum cryptography

    Lesson coming

Sample questions

Try each one before opening the answer. Every option is explained, with the AWS documentation page that proves it.

Question 1 · choose 2

An IAM role in an analytics account must read objects from an S3 bucket in a data account. The bucket uses SSE-KMS with a customer managed key in the data account, and the data must stay encrypted with that key. The bucket policy already allows the role s3:GetObject, but every read fails with an access denied error from AWS KMS. Which changes are required? (Choose TWO.)

  1. ACreate an interface VPC endpoint for AWS KMS in the analytics account so that the role's decrypt calls reach the key
  2. BAdd a statement to the key policy in the data account that allows the analytics account role to call kms:Decrypt
  3. CAdd kms:Decrypt for the analytics role to the bucket policy so that S3 passes the permission on to AWS KMS
  4. DRe-encrypt the objects with the AWS managed key for Amazon S3 and share that key with the analytics account
  5. EAdd an IAM policy to the role in the analytics account that allows kms:Decrypt on the key's ARN
Show the answer and why
  • ACreate an interface VPC endpoint for AWS KMS in the analytics account so that the role's decrypt calls reach the key

    Incorrect

    An interface endpoint gives a private network path to AWS KMS, but the access denied error comes from AWS KMS itself: the calls already reach it and fail authorization, which only the key policy and IAM policies can grant.

  • BAdd a statement to the key policy in the data account that allows the analytics account role to call kms:Decrypt

    Correct

    Cross-account use of a KMS key requires the key policy in the owning account to give the external account or its role permission to use the key.

  • CAdd kms:Decrypt for the analytics role to the bucket policy so that S3 passes the permission on to AWS KMS

    Incorrect

    Permission to use a KMS key comes from the key policy and IAM policies. A bucket policy controls access to the bucket and its objects, not to the KMS key.

  • DRe-encrypt the objects with the AWS managed key for Amazon S3 and share that key with the analytics account

    Incorrect

    The data must stay on the customer managed key. Besides, an AWS managed key can be used only by principals in its own account, its permissions cannot be managed, and resources encrypted under it cannot be shared with other accounts.

  • EAdd an IAM policy to the role in the analytics account that allows kms:Decrypt on the key's ARN

    Correct

    The external account must also delegate the key policy permission to its role with an IAM policy. Neither the key policy nor the IAM policy alone is enough.

Cross-account KMS needs both halves: the key policy says who can use the key, and the IAM policy in the other account says who does. The bucket policy only covers S3.

Question 2 · choose 1

A new public website will be served by an internet-facing Application Load Balancer in eu-west-1, and its domain is hosted in Route 53 in the same account. Visitors' browsers must trust the certificate without any client setup. Last year a certificate on another site expired because its renewal request went to a mailbox nobody read, so renewals must now complete with no human action at all, and no employee may ever handle private key material. Which approach meets these requirements?

  1. ARequest a public ACM certificate validated by email to the domain's administrative addresses and associate it with the load balancer
  2. BRequest a public ACM certificate with DNS validation, create its CNAME record in Route 53, and associate it with the load balancer
  3. CBuy a certificate from a public CA, import it into ACM, and alarm on the days remaining before it expires
  4. DUpload a certificate from a public CA to IAM as a server certificate and select it on the load balancer's HTTPS listener
Show the answer and why
  • ARequest a public ACM certificate validated by email to the domain's administrative addresses and associate it with the load balancer

    Incorrect

    A public ACM certificate is trusted by browsers and needs no key handling. With email validation, though, ACM sends validation emails again when renewal is due, and someone must follow the link, which is exactly the step that failed last year.

  • BRequest a public ACM certificate with DNS validation, create its CNAME record in Route 53, and associate it with the load balancer

    Correct

    ACM renews Amazon-issued certificates automatically when they use DNS validation, the private key stays in ACM, and a public certificate is trusted by browsers. Keeping the validation CNAME in Route 53 lets each renewal complete without anyone acting.

  • CBuy a certificate from a public CA, import it into ACM, and alarm on the days remaining before it expires

    Incorrect

    Imported certificates work with integrated services, but ACM does not renew them. Someone must obtain a new certificate and reimport it, which means handling the private key.

  • DUpload a certificate from a public CA to IAM as a server certificate and select it on the load balancer's HTTPS listener

    Incorrect

    AWS advises using IAM as a certificate manager only in Regions that ACM does not support. The certificate comes from an external provider, so renewal and key handling stay manual.

Browser trust, renewal with no human step and no key handling decide it. Imported and IAM certificates leave renewal and keys with staff. A public ACM certificate meets the trust and key constraints either way, but only DNS validation lets ACM renew it automatically; email validation needs someone to approve each renewal.

Question 3 · choose 1

A bank's cryptography board requires its applications to use a post-quantum key exchange whenever they call AWS KMS. The board wants evidence, for each KMS API call an application makes, of whether the connection used such a key exchange, and an alert when an application calls KMS without one. Which approach meets these requirements?

  1. ATurn on VPC Flow Logs for the application subnets and inspect the TLS handshakes recorded in the flow log records for each call
  2. BAdd a key policy condition on the KMS keys that requires a post-quantum TLS connection
  3. CRead tlsDetails.keyExchange in each KMS call's CloudTrail event and alarm with a metric filter when it lacks MLKEM
  4. DRun an AWS Config managed rule that evaluates whether the KMS keys use a post-quantum key spec
Show the answer and why
  • ATurn on VPC Flow Logs for the application subnets and inspect the TLS handshakes recorded in the flow log records for each call

    Incorrect

    Flow logs record IP traffic metadata such as addresses, ports and bytes. They do not show the TLS key exchange algorithm.

  • BAdd a key policy condition on the KMS keys that requires a post-quantum TLS connection

    Incorrect

    The documented way to verify and monitor hybrid post-quantum TLS for KMS is the CloudTrail record of each call, not a key policy condition.

  • CRead tlsDetails.keyExchange in each KMS call's CloudTrail event and alarm with a metric filter when it lacks MLKEM

    Correct

    CloudTrail records the TLS key exchange of each KMS call; a hybrid post-quantum connection shows an algorithm such as X25519MLKEM768. A CloudWatch metric filter can alert on calls without it.

  • DRun an AWS Config managed rule that evaluates whether the KMS keys use a post-quantum key spec

    Incorrect

    Key specs describe the key material of a KMS key. They say nothing about the TLS connection that an application used to call KMS.

Hybrid post-quantum TLS protects KMS API traffic against harvest now, decrypt later. CloudTrail's tlsDetails.keyExchange field proves which key exchange each call used, and a metric filter can flag calls without ML-KEM.

Question 4 · choose 1

A medical device maker signs firmware images in a build pipeline on AWS. A new regulation requires signatures that stay secure against future quantum computers. Devices in hospitals verify each image offline before installing it and hold no AWS credentials, and the private signing key must never leave AWS KMS unencrypted. Which approach meets these requirements?

  1. ACreate an asymmetric KMS key with an RSA_4096 key spec for signing, and embed its public key in the devices
  2. BCreate an ML-DSA asymmetric KMS signing key, sign in the pipeline, and embed the downloaded public key in the devices
  3. CCreate an HMAC KMS key, generate a MAC for each image, and have the devices call the KMS VerifyMac operation
  4. DCreate an asymmetric KMS key with an ECC_NIST_P521 key spec, which offers the largest elliptic curve key that KMS supports
Show the answer and why
  • ACreate an asymmetric KMS key with an RSA_4096 key spec for signing, and embed its public key in the devices

    Incorrect

    Devices could verify offline with the public key, but RSA is a classic algorithm, not one designed to resist quantum computers.

  • BCreate an ML-DSA asymmetric KMS signing key, sign in the pipeline, and embed the downloaded public key in the devices

    Correct

    ML-DSA is the post-quantum signature algorithm from FIPS 204. KMS keeps the private key inside the service, and the public key can be downloaded so that devices verify signatures without calling AWS.

  • CCreate an HMAC KMS key, generate a MAC for each image, and have the devices call the KMS VerifyMac operation

    Incorrect

    VerifyMac is an AWS API call, so devices without AWS credentials and network access could not verify images offline.

  • DCreate an asymmetric KMS key with an ECC_NIST_P521 key spec, which offers the largest elliptic curve key that KMS supports

    Incorrect

    A larger elliptic curve is still a classic algorithm and does not give the quantum resistance that the regulation requires.

For quantum-resistant digital signatures, AWS KMS offers ML-DSA key specs. The private key never leaves KMS unencrypted, and the public key can be exported for offline verification.

Question 5 · choose 1

An insurer's on-premises applications fetch database credentials from AWS Secrets Manager over the internet many times a day. A risk review warns that an adversary could record this TLS traffic now and decrypt it years later with a quantum computer, and the credentials protect records that must stay confidential for 25 years. The applications use a recent AWS SDK, and the insurer wants the smallest change. What should a solutions architect recommend?

  1. ARotate the secrets every day so that any credential recorded in transit stays valid only for a short time after capture
  2. BRe-encrypt the secrets with an asymmetric KMS key that uses an ML-DSA key spec
  3. CTurn on hybrid post-quantum TLS with ML-KEM in the SDK's HTTP client for the calls to Secrets Manager
  4. DRequire TLS 1.3 as the minimum protocol version in the SDK configuration of every application
Show the answer and why
  • ARotate the secrets every day so that any credential recorded in transit stays valid only for a short time after capture

    Incorrect

    Rotation limits how long a stolen credential works, but the recorded traffic would still be decryptable later, and the data behind the credentials must stay confidential for decades.

  • BRe-encrypt the secrets with an asymmetric KMS key that uses an ML-DSA key spec

    Incorrect

    ML-DSA keys sign and verify. They do not encrypt secrets, and the key used at rest does not change how the TLS session key is exchanged.

  • CTurn on hybrid post-quantum TLS with ML-KEM in the SDK's HTTP client for the calls to Secrets Manager

    Correct

    Secrets Manager endpoints support hybrid post-quantum key exchange that combines ECDH with ML-KEM, which protects recorded TLS traffic against harvest now, decrypt later.

  • DRequire TLS 1.3 as the minimum protocol version in the SDK configuration of every application

    Incorrect

    TLS 1.3 alone still uses a classic key exchange, which is the part a future quantum computer could break.

Long-lived confidentiality of data in transit calls for post-quantum key exchange. AWS KMS, ACM and Secrets Manager endpoints support hybrid ML-KEM TLS, which supported AWS SDKs can turn on in their HTTP client.

Question 6 · choose 1

A customer managed KMS key protects payroll data in S3 and Aurora. Its key policy has no statement that enables IAM policies; one statement gives the security team's KeyAdmin role kms:* and another lets the payroll application role use the key. An audit found that KeyAdmin can therefore decrypt payroll data. Auditors require that KeyAdmin keep managing the key, such as its policy, rotation and deletion schedule, without any cryptographic use, that only the application role can encrypt and decrypt, and that the key ID and existing ciphertext stay unchanged. Which change meets these requirements?

  1. ARewrite the key policy to give KeyAdmin only key administration actions and the application role only cryptographic actions
  2. BRemove the application role from the key policy and give it a grant for Encrypt and Decrypt on the key instead
  3. CRe-encrypt the payroll data with the AWS managed keys for Amazon S3 and Amazon RDS and retire the customer managed key
  4. DKeep the policy, and alarm on CloudTrail records of Decrypt calls made by KeyAdmin so that the security team investigates each one
Show the answer and why
  • ARewrite the key policy to give KeyAdmin only key administration actions and the application role only cryptographic actions

    Correct

    Key policies separate key administrators, who manage the key, from key users, who use it in cryptographic operations, in different statements. Because the policy does not enable IAM policies, the key policy alone decides access, and the key and its ciphertext stay as they are.

  • BRemove the application role from the key policy and give it a grant for Encrypt and Decrypt on the key instead

    Incorrect

    Grants let principals use a KMS key in cryptographic operations, often for temporary access, without editing the key policy. They do not take anything away from KeyAdmin, which keeps kms:* and can still decrypt.

  • CRe-encrypt the payroll data with the AWS managed keys for Amazon S3 and Amazon RDS and retire the customer managed key

    Incorrect

    Nobody can manage the lifecycle or permissions of an AWS managed key, so KeyAdmin could not manage rotation or deletion, and re-encrypting changes the key and the ciphertext.

  • DKeep the policy, and alarm on CloudTrail records of Decrypt calls made by KeyAdmin so that the security team investigates each one

    Incorrect

    CloudTrail records KMS API calls, so an alarm would reveal decryption by KeyAdmin after the fact. It detects rather than prevents, so KeyAdmin would still be able to decrypt payroll data.

The constraints are management without use, use without management, and an unchanged key. Grants add use without removing KeyAdmin's kms:*, an alarm only detects decryption, and AWS managed keys cannot be managed at all and force re-encryption. Splitting the key policy into a key administrators statement and a key users statement meets every constraint.

Question 7 · choose 1

A regulator requires an insurer's encryption keys for one workload to stay in a key manager that the insurer controls outside AWS. The insurer wants to know whether the workload can still use AWS services that encrypt with KMS. Which capability addresses this gap?

  1. AA CloudHSM cluster in the workload's VPC as a custom key store
  2. BAWS managed keys for each service
  3. CCustomer managed keys in KMS with automatic rotation turned on
  4. DAn AWS KMS external key store backed by the insurer's key manager
Show the answer and why
  • AA CloudHSM cluster in the workload's VPC as a custom key store

    Incorrect

    CloudHSM runs inside AWS, not in the insurer's own key manager.

  • BAWS managed keys for each service

    Incorrect

    AWS managed keys are kept inside AWS KMS.

  • CCustomer managed keys in KMS with automatic rotation turned on

    Incorrect

    Rotation changes key material but keeps it inside AWS KMS.

  • DAn AWS KMS external key store backed by the insurer's key manager

    Correct

    External key stores protect AWS resources with keys outside of AWS, in an external key management system that you control.

Keys that must stay outside AWS are used through an external key store.

Practice domain 2 →Practice all domains →