Prepare and test your skills

Prepare and test your skills

Infrastructure impact on business agility and sustainability describes how a worldwide cloud foundation helps organizations move fast while using less energy. Google Cloud's global infrastructure (opens in a new tab) is a network of data centers, fiber, undersea cables, and points of presence (PoPs) connected by a fast, reliable network. This foundation lets organizations scale quickly, place workloads close to users, and move work from older on-premises data centers to modern, efficient cloud facilities. Business agility and sustainability are linked because efficient workloads use fewer resources, cost less, and produce fewer emissions.
Global infrastructure enables business agility by giving applications a fast, redundant, and scalable foundation for change. Google Cloud's network uses its own fiber, publicly available fiber, undersea cables, and PoPs, so customer traffic can stay within Google's network until it is close to its destination. This reduces hops across the public internet, which lowers latency and limits opportunities for data to be attacked or intercepted. The network also supports security controls (opens in a new tab) such as firewalls, access control lists, Google Front End (GFE) servers, and distributed denial-of-service (DDoS) protection.
The following infrastructure features directly support rapid scaling and reliable operations:
Sustainability value of migrating to the cloud is the reduction in energy use and emissions from moving workloads to modern cloud data centers. Migrating on-premises workloads (opens in a new tab) to Google Cloud can reduce energy usage and associated emissions by 1.4 to 2 times compared with typical on-premises deployments. Cloud data centers are custom-designed for high power usage effectiveness (PUE), while older on-premises data centers often lack the scale to justify advanced cooling and power distribution. Google Cloud's carbon-neutral infrastructure, renewable energy investments, climate-conscious cooling, and hardware optimization support carbon-neutral data center operations.
Exam tip: Google is responsible for sustainability of Google Cloud, which covers data centers, infrastructure, and core services. Customers are responsible for sustainability in the cloud, which means optimizing their own workloads.
Customers improve sustainability in the cloud by managing resources efficiently. Key actions include right-sizing resources, using serverless services that scale to zero, and managing data lifecycles effectively. These actions reduce resource usage, cloud costs, and the carbon footprint of the workloads.
Sustainable by design means building systems that optimize for sustainability while also improving performance, cost, security, resilience, and user experience. An efficient application completes tasks faster, so compute resources run for a shorter time and consume fewer kilowatt-hours (kWh). Because cloud cost depends on resource usage, actions that save energy also save money. Strong security also helps sustainability by preventing attacks that cause traffic spikes and runaway compute cycles.
The relationships between sustainability and business outcomes work as follows:
A concept map with Sustainable by Design at the center linking performance, cost, security and resilience, and user experience to lower energy use, cost, and emissions.
Tradeoffs and decision criteria for global distribution balance higher availability and lower latency against higher data transfer cost and complexity. Google Cloud charges for traffic in three ways, and designing for high availability across zones or regions can increase cost. Smaller recovery time objective (RTO) (opens in a new tab) and recovery point objective (RPO) values, which mean faster recovery and less data loss, typically cost more and add administrative complexity. Google Cloud can lower the cost of meeting RTO and RPO compared with on-premises options because it provides a global network, redundancy, scalability, and security.
The three traffic types and their charging rules are:
| Traffic type | Direction | Charging |
|---|---|---|
| Ingress traffic | Enters Google Cloud from the public internet, on-premises locations, or other clouds | Free for most services; some internet-facing services such as Cloud Load Balancing, Cloud CDN, and Google Cloud Armor charge by ingress volume |
| Egress traffic | Leaves Google Cloud in any way | Charges apply to many services, including Compute Engine, Cloud Storage, Cloud SQL, and Cloud Interconnect |
| Regional and zonal traffic | Crosses regional or zonal boundaries inside Google Cloud | Can be subject to bandwidth charges; for example, sending traffic to a database replica in another zone is subject to cross-zonal charges |
Organizations can use resource location constraints to limit workloads to pre-approved regions and choose low-carbon regions to support sustainability goals. The most cost-efficient region might not have the lowest carbon footprint or lowest latency, so the best choice depends on which requirement matters most.
Tools and practices for sustainable infrastructure help teams design energy-efficient and carbon-aware workloads. The Well-Architected Framework sustainability pillar (opens in a new tab) provides core principles for this work. AI can also turn large datasets into actionable insights and help organizations move from reactive compliance to proactive optimization.
The sustainability principles cover region choice, workload efficiency, data management, and culture:
Google products that support sustainability goals include Google Earth AI for analyzing environmental changes and supply chain impacts, WeatherNext for weather forecasting and climate risk analytics, Geospatial insights with Google Earth for site selection and resource planning, and Google Maps routes optimization for reducing fuel consumption and transportation emissions. For example, Google Cloud and TELUS partnered to migrate workloads to carbon-neutral infrastructure and use data analytics for smart-city technology that reduces traffic congestion and carbon emissions.
Hierarchical structure of regions and zones is the way Google Cloud organizes locations from worldwide network to region to zone. Regions (opens in a new tab) are geographic areas that contain multiple zones, and zones sit inside regions as single deployment areas. A zone is a single deployment area for resources such as Compute Engine virtual machines, while a region supports regional and multi-regional resources. This structure lets an organization choose how widely to spread an application, from one zone for simple testing to multiple zones and regions for production resilience and compliance.
What regions and zones mean for resources is the scope where a resource runs and where its data is stored. Zonal resources, such as Compute Engine (opens in a new tab) virtual machines and zonal Persistent Disk volumes, live in one zone and depend on that zone being available. Regional resources, such as App Engine applications, regional managed instance groups, regional Persistent Disk volumes, and Filestore regional instances, spread across zones within one region. Multi-regional resources, such as Cloud Storage, Datastore, Firestore, and Spanner (opens in a new tab) in multi-region configurations, store data redundantly in at least two separate geographic locations.
How zones provide high availability is the use of separate failure boundaries inside the same region so an application keeps running when one zone fails. An application can use regional resources that automatically replicate or distribute work across zones, so traffic flows from users through global load balancing to healthy zones when one zone fails. If it uses zonal resources instead, the customer must manage its own compute and storage redundancy across zones or across regions. Control planes are designed with the goal that one region is not impacted if other regions become unavailable, and Google validates this with periodic disconnect testing where network connectivity to an internal-only region is physically disconnected.
Storage choices show how zone-level protection works in practice. Regional Persistent Disk volumes provide synchronous replication across two zones in a region, so a write is copied to both zones before it completes. Data in Cloud Storage regional buckets is replicated synchronously across the zones in the region, while data in a Filestore regional instance is replicated synchronously across three zones in the region. These synchronous copies keep data available without a separate failover step when a single zone has a problem.
| Resource scope | Where data is kept | Failure it is designed to withstand |
|---|---|---|
| Zonal Persistent Disk volume | One zone | Requires customer-managed copy to survive zone loss |
| Regional Persistent Disk volume | Two zones in one region, synchronously replicated | Zone outage within the region |
| Cloud Storage regional bucket and Filestore regional instance | Zones in one region, synchronously replicated | Zone outage within the region |
| Cloud Storage dual-region or multi-region bucket | At least two geographic locations; metadata written synchronously, data replicated asynchronously | Extended loss of an entire region |
How regions provide fault tolerance and disaster recovery (opens in a new tab) is the use of larger geographic boundaries so an application can withstand the extended loss of an entire region. For data, an organization can use managed multi-regional storage services, snapshot zonal or regional data to a multi-regional resource, or manage its own data replication to one or more other regions. For compute, an organization can use zonal or regional resources but manually or automatically bring up the application in another region, referring to copies of primary data if the data is not already in a managed multi-regional resource. Cross-region networking carries explicitly configured backups and replication to other regions, and global load balancing, DNS, and failover between regions depend on physical networking infrastructure.
Exam tip: Use multiple zones in one region to survive a zone outage, and use multiple regions or multi-regional storage to survive the extended loss of an entire region.
Recovery planning ties region design to cost through two metrics: recovery time objective (RTO), the maximum acceptable time an application can be offline, and recovery point objective (RPO), the maximum acceptable time window for which data might be lost. Smaller RTO and RPO values generally mean higher cost and greater administrative complexity because distribution, replication, and failover must be managed across more locations. For dual-region buckets, turbo replication provides a recovery point objective of 15 minutes for replication across the region pair, which helps planners match a replication choice to a required RPO.
How geographic distribution supports data residency and compliance is the ability to control where data is stored and processed to meet regulatory requirements such as HIPAA or GDPR. Data locality for Google Cloud services is governed by the terms of service, including service-specific terms, and different services have different regionality that must be considered when choosing a deployment target. An organization can use resource location constraints to help ensure that workloads can only be deployed in pre-approved regions. When regional or zonal storage is required, Google Cloud recommends replicating data to another region or snapshotting it to a multi-regional storage resource for disaster recovery.
Choosing where to deploy is the decision of which region and zone spread best balances performance, cost, availability, and policy requirements. The Google Cloud foundation (opens in a new tab) provides global reach, but each region differs in features, price, latency, and carbon impact. Points of presence help with performance and security by letting customer traffic travel within the Google network until it is close to its destination. Together, region choice, zone spread, and storage scope determine how an application balances speed, cost, resilience, and control over data location.
Documented factors for region selection include:
Technical advantages of the global network are faster delivery, stronger security, and higher availability from a private worldwide backbone. The global network (opens in a new tab) connects Google data centers around the world using owned fiber, public fiber, and undersea cables. Customer traffic enters this network early and stays on it until it is close to its destination, instead of traveling long distances over the public internet. This design delivers linked advantages for digital transformation that can scale like consumer products such as Search and Gmail, even during large traffic spikes.
Private network and points of presence (PoPs) are the entry points where user traffic moves from the public internet onto Google's private backbone. Regions (opens in a new tab) connect through this backbone, and PoPs serve as global peering locations linked to most internet service providers (ISPs) in the world. Traffic flows from the user device to a nearby PoP and then travels within the Google network to the destination data center. This path limits the number of hops, which are the intermediate steps data takes across the public internet between a customer ISP and a data center.
Fewer hops matters for both speed and security. Each additional hop on the public internet adds delay and creates another opportunity for data to be attacked or intercepted. By keeping traffic on its private backbone for most of the journey, Google reduces latency for global users and limits exposure to bad actors. Access to Google Cloud APIs also depends on these PoPs working together with globally distributed load balancers and DNS.
Low latency and high-speed delivery mean data travels quickly with minimal delay and consistent performance worldwide. The physical foundation is an IP data network built from Google-owned fiber, publicly available fiber, and undersea cables that connects regions across the globe. On top of that foundation, the backbone uses advanced software-defined networking (SDN) and edge-caching services to deliver fast, consistent, and scalable performance. Software-defined networking virtualizes the physical network layer, so network behavior is controlled by software rather than by manually configuring each router and switch.
Exam tip: Google's private backbone keeps traffic off the public internet for most of the trip, so fewer public hops means lower latency and less chance of interception.
This software control allows Google to route traffic efficiently, balance load across paths, and cache content close to users at the network edge. The result is high-speed data delivery that scales automatically as demand changes. Edge caching places frequently used content near users, so repeat requests do not need to travel all the way to the origin data center.
Security and availability built into the network come from defense in depth and redundancy of everything. Defense in depth means multiple protective layers guard traffic as it flows from users through Google Front End (GFE) (opens in a new tab) servers to Google services. Traffic is routed through GFE servers to help detect and stop malicious requests and distributed denial-of-service (DDoS) attacks. Redundancy of everything means no service depends on a single server, data center, or network connection, and data centers are geographically distributed to minimize the effects of regional disruptions such as natural disasters or local outages.
The network enforces several peer controls to protect that flow:
Availability relies on redundant components and distributed locations working together:
This combination helps deliver highly available, low-latency services worldwide, and it supports protections such as Project Shield, which provides free DDoS protection for news, human rights, and election-monitoring websites. Secure networking design (opens in a new tab) ties these controls together so traffic stays protected from the network edge to the destination service.
Google Cloud's global network connects data centers with owned fiber, publicly available fiber, undersea cables, and points of presence (PoPs) so customer traffic stays on Google's private network until it is close to its destination. Fewer hops across the public internet lowers latency and limits opportunities for data to be attacked or intercepted, while Google Front End (GFE) servers, firewalls, access control lists, and distributed denial-of-service (DDoS) protection help detect and block malicious requests.
Use multiple zones in one region to survive a zone outage and use multiple regions or multi-regional storage to survive the extended loss of an entire region. Regional Persistent Disk volumes replicate data synchronously across two zones in a region and Filestore regional instances replicate synchronously across three zones in a region, while Cloud Storage dual-region or multi-region buckets store data redundantly in at least two geographic locations.
Migrating on-premises workloads to Google Cloud can reduce energy usage and associated emissions by 1.4 to 2 times compared with typical on-premises deployments because cloud data centers are custom-designed for high power usage effectiveness (PUE) with carbon-neutral operations, renewable energy investments, climate-conscious cooling, and hardware optimization. Google is responsible for sustainability of Google Cloud including data centers and infrastructure, while customers are responsible for sustainability in the cloud by right-sizing resources, using serverless services that scale to zero, and managing data lifecycles effectively.
A global retail organization is migrating its legacy on-premises data centers to Google Cloud as part of its digital transformation strategy. The executive leadership team has established two primary requirements for the new cloud environment:
Which architectural approach on Google Cloud's global infrastructure should the organization implement to meet these requirements?
Keep variable compute workloads in on-premises data centers and use Google Cloud Coldline storage strictly for archiving historical data
Provision fixed-capacity virtual machines across multiple regions in an active-passive configuration to maintain continuous standby compute power
Deploy workloads in Google Cloud regions with high Carbon-Free Energy (CFE) scores and configure automated horizontal scaling to match resource capacity with real-time demand
Scale virtual machines vertically to their maximum capacity and rely solely on Google Cloud's data center cooling efficiency to reduce emissions
Last updated