Prepare and test your skills

Prepare and test your skills

The global private network (opens in a new tab) is Google's worldwide IP data network built from Google-owned fiber, public fiber, and undersea cables. Customer traffic enters this private network as early as possible instead of traveling long distances over the public internet. This design links three benefits together: less exposure to interception, faster delivery, and continued operation during regional disruptions.
Google's private fiber and subsea cable network is the physical foundation that moves traffic between users and Google data centers around the world. In other setups, data travels between devices over the public internet in paths called hops, and the number of hops depends on the route between the customer's internet service provider (ISP) and the data center. Because Google's global network connects directly to most ISPs in the world, it limits the number of hops across the public internet. Each hop that is avoided removes another opportunity for data to be attacked or intercepted.
Customer traffic leaves its ISP after one hop and enters Google's private fiber and subsea network at the edge, passing Google Front End filtering before traveling to distributed data centers.
The security protections on Google's global network come from keeping traffic off the public internet and defending the private network in layers. Defense in depth means several independent controls protect the network, so traffic must pass authorized checks at more than one layer. The layers work together in this way:
Traffic flows from customer devices to GFE servers at the edge of Google's network, where harmful requests are filtered before they reach production systems. The same global infrastructure also runs Project Shield, which provides free, unlimited protection against DDoS attacks used to censor information for news, human rights, and election-monitoring websites.
Exam tip: Fewer public internet hops means less exposure — Google limits hops by linking its private network directly to most ISPs, so traffic enters Google's protected network early.
Low-latency and highly available solutions on Google Cloud depend on the private IP data network combined with redundancy at every layer of the platform. Low latency means data arrives quickly, while high availability means services stay reachable even when a component fails. Google designs its platform for redundancy of everything, so no solution depends on a single server, data center, or network connection:
Google's data centers are geographically distributed to minimize the effects of regional disruptions on global products. When a natural disaster or local outage affects one region, traffic can use other network paths and data centers to keep services highly available and responsive across the globe.
Purpose-built data centers and custom servers improve efficiency by matching facilities, power, cooling, and hardware to cloud workloads. Google designs and builds its own data centers, servers, network equipment, and some security chips, so it can control energy use, cooling, performance, and security together. This approach supports strong power usage effectiveness (PUE), lower cost, and lower environmental impact through continuous investment in sustainability (opens in a new tab).
Google data centers are custom-designed facilities built for power usage effectiveness (PUE), a measure of how efficiently a data center uses electricity. Older on-premises data centers often lack the scale needed to invest in advanced cooling and power distribution systems. Google data centers use several operational controls to keep services running while limiting environmental impact:
Google was the first major internet services company to get external certification of its high environmental, workplace safety, and energy management standards throughout its data centers. For example, it obtained voluntary ISO 50001 certifications for data centers in Europe.
Custom servers and network equipment are purpose-built to maximize performance, cooling, and power efficiency. Unlike most commercially available hardware, Google servers omit unnecessary components such as video cards, chipsets, or peripheral connectors, which can introduce vulnerabilities and waste resources. Google vets component vendors and audits the security properties of components. It also designs custom chips such as Titan, which help securely identify and authenticate legitimate Google devices at the hardware level, including the code those devices use to boot up.
Server resources are dynamically allocated. This allocation gives flexibility for growth and lets Google adapt quickly and efficiently to customer demand by adding or reallocating resources. Proprietary software continually monitors systems for binary-level modifications. Automated self-healing mechanisms monitor and remediate destabilizing events, receive incident notifications, and slow potential compromises on the network.
The hardware lifecycle is how Google tracks, retires, and disposes of data center equipment without exposing customer data. Google tracks the location and status of equipment in data centers using barcodes and asset tags. Metal detectors and video surveillance help prevent equipment from leaving the data center floor without authorization. If a component fails a performance test at any point during its lifecycle, it is removed from inventory and retired.
Storage devices, including hard drives, solid-state drives, and non-volatile dual inline memory modules (DIMMs), use technologies such as full disk encryption (FDE) and drive locking to protect data at rest. When a storage device is retired, authorized individuals verify that it is sanitized and perform a multiple-step verification process to ensure it contains no data. If a device cannot be erased, it is physically destroyed using a shredder that breaks it into small pieces, which are then recycled at a secure facility.
Migrating on-premises workloads to cloud infrastructure can reduce energy use and associated emissions by 1.4 to 2 times compared with typical on-premises deployments. The comparison below shows why scale and custom design affect sustainability and cost.
| Environment | Energy and emissions | Cooling and power design |
|---|---|---|
| Typical on-premises data center | Higher energy use and emissions | Often lacks the scale to justify advanced cooling and power distribution systems |
| Google Cloud data center | Reduces energy use and emissions by 1.4 to 2 times when workloads migrate | Modern, custom-designed facility built for power usage effectiveness (PUE) |
Under the shared responsibility model, Google is responsible for sustainability of Google Cloud, meaning the energy efficiency and water stewardship of its data centers, infrastructure, and core services. Customers are responsible for sustainability in the cloud, meaning they optimize workloads to be energy efficient. For example, customers can right-size resources, use serverless services that scale to zero, and manage data lifecycles effectively.
Exam tip: Google is responsible for sustainability of the cloud, while customers are responsible for sustainability in the cloud.
Custom hardware and security integration means Google designs its own data centers, purpose-built servers, networking equipment, and security hardware so protection is built into the physical infrastructure. The Titan security chip (opens in a new tab) is a custom-designed component that acts as a hardware root of trust. This root of trust gives Google's global infrastructure a trusted starting point for checking the integrity of the boot process.
A hardware root of trust (RoT) is a protected part of a system that is trusted first because its security functions are built into hardware. The Titan security chip provides this anchor for Google infrastructure. Because the hardware root of trust is trusted before the operating system or applications start, later security checks can depend on it. This relationship is one-way: software trust depends on the hardware root, not the other way around.
The Titan security chip helps ensure the integrity of the boot process by checking that firmware and early boot components are the expected, unmodified versions before they are allowed to run. The boot process follows a chain of trust in a fixed order:
A left-to-right process flow showing Titan starting as hardware root of trust, verifying first firmware then each next stage, passing trust forward, and continuing only if checks succeed or stopping if they fail.
This chain means an altered boot component cannot silently take control, because the earlier trusted stage will not pass trust to it.
Google's global infrastructure includes many data centers and servers in different regions. Using the same custom Titan security chip and hardware root of trust across that infrastructure gives each machine a consistent, hardware-based starting point for boot integrity. This consistency helps security teams trust that servers in any location begin from a known good state. It also supports vertically integrated infrastructure: Google controls the design of its data centers, purpose-built servers, networking, and security hardware, so security, performance, and reliability can be designed together rather than assembled from unrelated parts.
Exam tip: A hardware root of trust is trusted first and cannot be replaced by software trust; the boot process builds a chain of trust from that hardware anchor.
Google's global private network connects directly to most internet service providers (ISPs) so traffic enters private fiber and subsea cables early and avoids extra hops across the public internet. Each avoided hop removes another opportunity for data to be attacked or intercepted, while layered defenses including authorized services and protocols, firewalls, access control lists, and Google Front End (GFE) servers filter malicious requests before they reach production systems.
Power usage effectiveness (PUE) is a measure of how efficiently a data center uses electricity. Google data centers are custom-designed facilities built for power usage effectiveness (PUE) with redundant power, cooling systems that maintain a constant operating temperature, and fire protection, and migrating workloads to them can reduce energy use and emissions by 1.4 to 2 times compared with typical on-premises deployments.
The Titan security chip acts as a hardware root of trust that verifies firmware and early boot components are the expected, unmodified versions before they are allowed to run. Each verified stage passes trust to the next stage in order, so an altered component cannot silently take control because the earlier trusted stage will not pass trust to it.
Sustainability of Google Cloud is Google's responsibility for the energy efficiency and water stewardship of its data centers, infrastructure, and core services. Sustainability in the cloud is the customer's responsibility to optimize workloads to be energy efficient, for example by right-sizing resources, using serverless services that scale to zero, and managing data lifecycles effectively.
An enterprise is planning to migrate workloads from its aging on-premises facilities to Google Cloud to reduce operational costs and meet corporate sustainability goals. The executive team wants to understand how Google's physical infrastructure inherently optimizes energy consumption, cooling, and resource utilization compared to traditional data centers.
How do Google's purpose-built data centers and custom servers achieve these operational and environmental benefits?
Last updated