Prepare and test your skills

Prepare and test your skills

Encryption for data at rest protects stored data by using cryptographic keys to make the data unreadable if storage is lost, stolen, or compromised. Default encryption (opens in a new tab) in Google Cloud protects customer data stored in Google production data centers with no customer action required. Organizations that need more control can manage their own keys, while still combining encryption with access control, data segmentation and residency, and auditing.
Google Cloud default encryption protects customer data at rest without customer configuration. Encryption is applied at the application layer, the storage device layer, or both layers, depending on the service. Keys are held in Keystore, Google Cloud's internal key management service, and are managed automatically by Google. Most services rotate these keys automatically, with a primary key version that encrypts new data encryption keys and a limited number of older versions that can still decrypt existing keys.
Customers cannot view or manage Keystore keys or review their usage logs, and data from multiple customers might share the same key encryption key. The cryptographic modules used are validated to FIPS 140-3 Level 1. Default encryption therefore removes the risk of storing unencrypted data by accident, but it does not give the customer control over key location, rotation, or destruction.
Exam tip: Google Cloud default encryption is not a feature of Cloud Key Management Service (Cloud KMS).
Customer-managed encryption keys (CMEKs) are keys you create and control in Cloud Key Management Service (Cloud KMS) (opens in a new tab) to protect resources in CMEK-integrated services. You control the key location, protection level, creation, access control, rotation, use, and destruction. CMEKs can be software-based or protected by hardware security modules to meet regulatory or compliance needs, such as a requirement to rotate keys regularly.
Cloud KMS Autokey can automatically provision and assign CMEKs to new resources. Using Autokey is simpler than creating keys yourself and is recommended when the keys it creates meet all requirements. CMEKs reduce risk from storage compromise because an attacker who obtains the encrypted data still cannot read it without access to the customer-controlled keys.
Customer-supplied encryption keys (CSEKs) are encryption keys that you provide to Google Cloud when needed for supported services. CSEKs are supported for Cloud Storage and Compute Engine, although CSEKs for Compute Engine are deprecated. With CSEKs, the key material lives in memory while in use, and Google does not permanently store your keys on its servers.
You must supply the key material with each request that needs it and remain responsible for retaining it. Using CSEKs might increase costs for Compute Engine or Cloud Storage. Choose CSEKs when the organization wants direct control over the key material itself, rather than letting Google store the keys.
Cloud External Key Manager (Cloud EKM) lets you encrypt Google Cloud data with keys that live in a third-party key management system outside Google's infrastructure. Cloud KMS works with Cloud EKM so workloads can still use cloud storage, compute, and analytics while the keys remain external. This design maintains separation between data at rest in Google Cloud and the encryption keys that can decrypt it.
Cloud EKM helps mitigate unauthorized access and storage compromise because control of the keys stays with the external system. If access to the external keys is removed or blocked, the encrypted data in Google Cloud cannot be decrypted.
The key management options for data at rest differ in who manages the keys, where key material lives, and how much lifecycle control the customer keeps.
| Option | Who manages keys | Where key material lives | Control and rotation |
|---|---|---|---|
| Google Cloud default encryption | Google manages keys in Keystore | Google infrastructure | Automatic rotation by most services; customer cannot view keys or usage logs |
| Customer-managed encryption keys (CMEKs) | Customer manages keys in Cloud KMS | Cloud KMS as software keys or HSM-protected keys | Customer controls location, protection level, access, rotation, use, and destruction; Autokey can automate provisioning |
| Customer-supplied encryption keys (CSEKs) | Customer provides key material | In memory; Google does not permanently store keys | Customer supplies keys when needed; Compute Engine support is deprecated |
| Cloud External Key Manager (Cloud EKM) | Customer manages keys in third-party system | Outside Google infrastructure | Customer controls keys through the external system while using Google services |
Encryption at rest reduces the risk of unauthorized access because stored data is unreadable without the correct keys. Default encryption provides this protection automatically for all customer data at rest. CMEKs, CSEKs, and Cloud EKM add stronger control by letting the organization decide where keys live, who can use them, when they rotate, and when they are destroyed. HSM-protected keys and externally managed keys can also satisfy stricter regulatory needs. These key controls work best together with access control, data segmentation and geographical residency, and auditing.
Encryption in transit (opens in a new tab) protects data while it moves between users, services, and regions. Data is encrypted before it leaves the sender, then decrypted and checked by the receiver, so interception during travel does not expose readable content. In Google Cloud this protection covers traffic from the internet to Google and traffic moving inside Google's network between virtual machines (VMs) and services.
Encryption in transit in Google Cloud combines three protections that work together on every connection. Authentication verifies the identity of each peer, whether a person or a service process, in a connection. Integrity prevents data from being altered while it travels from source to destination. Encryption uses cryptography to keep data confidential and unreadable while in transit.
Together these protections remove the need to trust the lower layers of the network. Even if traffic is intercepted between an end user and Google Cloud or between two internal services, an attacker cannot read it or quietly modify it.
Traffic routing in Google Cloud describes how a request flows from an end user on the internet to its destination and how services communicate inside Google. A Google Cloud service is a Google-provided modular service such as Cloud Storage, while a customer application is an application a customer builds and runs on services such as Cloud Run, Google Kubernetes Engine, or a Compute Engine VM. Requests to Google Cloud services and to customer applications that use load balancing flow through Google Front End (GFE), a globally distributed system that terminates incoming connections and routes traffic.
Google Front End (GFE) receives HTTP(S), TCP, and Transport Layer Security (TLS) connections at points of presence around the world. GFE provides distributed denial-of-service (DDoS) countermeasures and routes and load balances traffic to the correct service or application over Google's network. Inside Google, additional paths include VM to VM, VM to GFE, GFE to Google APIs and services, and Google Cloud service to Google Cloud service, and each path has its own encryption.
Encryption between end users and Google Cloud services protects requests sent over the internet with Transport Layer Security (TLS). When a client sends HTTPS, HTTP/2, or HTTP/3 traffic to a service such as Cloud Storage, traffic flows from the user to GFE and then from GFE over Google's network to the service. TLS in GFE is implemented with BoringSSL, a Google-maintained open-source implementation whose core, BoringCrypto, is validated to FIPS 140-3 Level 1.
GFE negotiates industry-standard cryptographic parameters with the client, including forward-secure key negotiation. For modern clients, GFE negotiates more modern protocols such as TLS or QUIC when possible. For older, less capable clients, GFE selects the strongest legacy option the client supports, so the connection still receives the best available protection.
Encryption between end users and customer applications protects traffic from the internet to applications hosted on Google Cloud. Traffic can flow directly to the external IP address of the VM that hosts the application, or it can flow through a load balancer. When an external Application Load Balancer or external proxy Network Load Balancer is used, the connection between the end user and the load balancer is encrypted with TLS by default when the client supports TLS.
GFE terminates those TLS connections using self-managed or Google-managed certificates, with free and automated certificates available for cases such as Firebase Hosting and Cloud Run custom domains. Protection can be extended beyond the load balancer by enabling encryption from the load balancer to the backend VM that hosts the application. For stronger two-way trust, mutual TLS (mTLS) authenticates both client and server, including through Cloud Service Mesh for workloads on Google Kubernetes Engine and Compute Engine using customer-managed certificates.
Exam tip: Connections between VMs that use internal IP addresses are automatically encrypted, while connections that use external IP addresses leave Google's network and are not automatically encrypted.
Secure service-to-service communication protects data moving between VMs and between Google production services inside Google's network. Google Cloud's virtual network encrypts, integrity-protects, and authenticates private IP traffic within the same Virtual Private Cloud (VPC) network or across peered VPC networks at the network layer. Each pair of communicating hosts establishes a session key over a control channel protected by Application Layer Transport Security (ALTS), uses that key to encrypt VM-to-VM traffic, and rotates session keys periodically.
Application Layer Transport Security (ALTS) is Google's mutual authentication and encryption system for traffic from GFE to a Google Cloud service and from one Google Cloud service to another. Each service holds credentials issued by Google's internal certificate authority (CA) and verifies the other service's credentials during remote procedure calls. ALTS then encrypts the traffic, by default with AES-128-GCM, and provides cryptographic integrity protection, so application traffic does not depend on the security of the network path.
The network layer is strengthened by the PSP security protocol (PSP), a transport-independent protocol that provides per-connection security with a unique encryption key for each connection. PSP supports TCP and non-TCP protocols such as UDP and supports offloading encryption to SmartNIC hardware when available. Google infrastructure (opens in a new tab) uses PSP to encrypt traffic within and between data centers, including traffic already protected by ALTS.
Google's global network reduces exposure during transmission by keeping traffic off the public internet as much as possible. In many other environments, data travels across the public internet through multiple hops between a customer internet service provider (ISP) and a data center, and each hop creates another chance for interception. Because Google's network connects to most ISPs in the world, traffic from the user reaches GFE quickly and then travels over Google's network, which limits public hops.
The network also uses defense in depth with multiple protective layers. Only authorized services and protocols that meet Google's security requirements can traverse the network, while anything else is automatically dropped. Firewalls and access control lists enforce segregation, traffic routed through GFE servers is checked to detect and stop malicious requests and DDoS attacks, logs are examined for exploitation, and access to network devices is limited to authorized employees.
Encryption for data in use and key management closes the gaps left by default protection in Google Cloud. Google Cloud encrypts customer data at rest and data in transit by default (opens in a new tab) with no action required, but data in use in memory is not encrypted by default. Confidential Computing protects data while it is being processed, and Cloud Key Management Service (Cloud KMS) controls the encryption keys that protect stored data.
Confidential Computing is the Google Cloud family of services that encrypts data in use while workloads run. Confidential Computing performs computation inside a Trusted Execution Environment (TEE) that provides hardware isolation and attestation. The TEE runs the workload in cryptographic isolation, which helps ensure confidentiality in a multi-tenant cloud, helps prevent unauthorized access or modification to applications and data while in use, and produces independently verifiable attestations about system state and code run.
Confidential Computing reduces the trust boundary around confidential data so fewer resources and people can access it in cleartext. Choose Confidential Computing when the organization handles sensitive and regulated data and needs verifiable security and privacy assurances for processing. Google Cloud provides Confidential Computing for different workload types, including virtual machines, containers, streaming analytics, data processing, and joint analysis.
Exam tip: Data at rest and data in transit are encrypted by default, while data in use in memory is not encrypted by default and requires Confidential Computing.
Cloud Key Management Service (Cloud KMS) is the Google Cloud service for creating, storing, rotating, and controlling encryption keys. Customer-managed encryption keys (CMEKs) created in Cloud KMS use envelope encryption to protect storage, compute, and big data workloads. Cloud KMS Autokey automates the provisioning and assignment of CMEKs, which simplifies consistent protection and regular rotation for regulatory or compliance needs.
The choice of key type determines where the key lives and how much control the customer keeps. Software keys are managed inside Google Cloud, hardware-backed keys add stronger protection, and external keys stay outside Google infrastructure while still allowing cloud compute and analytics on the encrypted data.
| Key option | Where keys are stored and managed | Choose when |
|---|---|---|
| Software-based CMEK in Cloud KMS | Keys are generated and stored as software in Cloud KMS | Basic customer control over generation, storage, and rotation is enough |
| HSM-protected CMEK with Cloud HSM | Keys are protected by hardware security modules in Cloud KMS | Regulatory or compliance rules require hardware protection and regular rotation |
| External key with Cloud External Key Manager (Cloud EKM) | Keys are stored and managed in a third-party system outside Google infrastructure | Separation between data at rest and encryption keys must be maintained |
Cloud External Key Manager (Cloud EKM) is the Cloud KMS capability that lets data be encrypted with keys stored outside Google infrastructure. Cloud EKM maintains separation between encrypted data at rest in Google Cloud and the keys that can decrypt it, while the data can still flow from Google storage to Google compute and analytics services for processing. Titan hardware security chips in Google servers support this model with secure key storage, root of trust, and signing authority.
Key Access Justifications adds approval control on top of hardware-backed and external keys. Key Access Justifications (opens in a new tab) provides a justification record every time an externally hosted key is used to decrypt data and requires Cloud KMS with Cloud HSM or Cloud KMS with Cloud EKM. The customer must approve access before Google personnel can decrypt data at rest, so no decryption can happen without a visible, reviewable reason.
Encryption protects data in all three states as it moves through Google Cloud. Data at rest is protected by default encryption or customer-controlled keys, data in transit is protected by TLS, ALTS, and network-layer encryption as traffic flows from users through Google Front End (GFE) to services and between services, and data in use is protected by Confidential Computing inside isolated hardware. Key management ties the states together by deciding who controls decryption, where keys live, and what approval is required to use them.
Google Cloud default encryption protects stored data automatically with keys held in Keystore that are managed and rotated by Google and cannot be viewed or managed by customers. Customer-managed encryption keys (CMEKs) are keys you create and control in Cloud Key Management Service (Cloud KMS) to protect resources in CMEK-integrated services, giving you control over key location, protection level, access, rotation, use, and destruction. Cloud KMS Autokey can automatically provision and assign CMEKs to new resources when the keys it creates meet all requirements.
You should use Confidential Computing when you need to encrypt data in use while workloads run, because data at rest and data in transit are encrypted by default while data in use in memory is not encrypted by default. Confidential Computing performs computation inside a Trusted Execution Environment (TEE) that provides hardware isolation and attestation to help prevent unauthorized access or modification while in use. Choose it when handling sensitive and regulated data that needs verifiable security and privacy assurances for processing.
Application Layer Transport Security (ALTS) is Google's mutual authentication and encryption system for traffic from Google Front End (GFE) to a Google Cloud service and from one Google Cloud service to another. Each service holds credentials issued by Google's internal certificate authority and verifies the other service's credentials during remote procedure calls, then encrypts traffic by default with AES-128-GCM with cryptographic integrity protection. This protection means application traffic does not depend on the security of the network path.
A financial services organization is migrating sensitive workloads to virtual machines on Google Cloud. To meet strict regulatory compliance standards, the organization must satisfy the following security requirements:
Which combination of Google Cloud security capabilities should the organization implement?
Last updated