Prepare and test your skills

Prepare and test your skills

Security Policy Implementation and Traffic Filtering is the practice of attaching security policies to load balancing so unwanted traffic is blocked before it reaches backend services. A client request first arrives at Google's edge, the Google Cloud Armor (opens in a new tab) security policy checks the request against allow and deny rules, and only permitted traffic flows forward to the backend. This edge-first order protects applications from unauthorized access and from large-scale Distributed Denial-of-Service (DDoS) (opens in a new tab) attacks that try to make services unavailable.
A Google Cloud Armor security policy is a set of rules that allows or denies incoming traffic based on characteristics of the request. The policy acts at the edge on client traffic arriving at Google Front Ends (GFEs) (opens in a new tab), which are Google's global entry points for user traffic. Because filtering happens at the GFE before backends are contacted, malicious or unauthorized requests never consume backend capacity.
Security policies support several types of traffic filters that can be used alone or together:
Integration with Cloud Load Balancing (opens in a new tab) means associating a Cloud Armor security policy with a load balancer backend service behind an external Application Load Balancer. Traffic flows from the client to the GFE, from the GFE through the security policy check, and then from the load balancer to the backend service only if the policy allows it. With complex URL maps that reference multiple backend services, the same policy can be reused across services or different policies can be attached to one or more backend services as needed.
Cloud Load Balancing also provides built-in network protections that work before and during this policy check. The Google frontend infrastructure terminates user traffic and automatically scales to absorb certain attacks, such as SYN floods, before they reach Compute Engine instances. Anycast-based load balancing uses a single anycast IP to front-end instances in all regions, directs traffic to the closest backend, and increases the surface area to absorb a DDoS attack by moving traffic to instances with available capacity in any region where backends are deployed.
The combination of Cloud Armor and Virtual Private Cloud (VPC) (opens in a new tab) firewall rules creates two separate control points for incoming traffic. Cloud Armor provides edge security at the GFE, while VPC firewall rules allow or deny traffic to and from the backends themselves. Ingress allow firewall rules must target the load-balanced backend virtual machines (VMs) and use the IP ranges of global external Application Load Balancers or classic Application Load Balancers as sources, so GFEs and health check systems can still communicate with backend VMs.
| Control | Where it acts | What it controls |
|---|---|---|
| Google Cloud Armor security policy | Client traffic to Google Front Ends (GFEs) | Which external clients can reach the load balancer, based on IP, geography, and list-based rules |
| VPC firewall rules | Traffic to and from backend VMs | Which load balancer and health check traffic can reach backends, blocking direct access that bypasses the load balancer |
A common deployment allows traffic only from trusted CIDRs to reach the load balancer while blocking direct access to backend instances. Only external traffic proxied through the load balancer with its associated security policy can reach the instances:
us-west1 region and another in the europe-west1 region.120.1.1.1 external IP address.Exam tip: A Cloud Armor security policy filters client traffic at the GFE, while VPC ingress allow firewall rules permit only load balancer and health check traffic to reach backend VMs.
DDoS Mitigation and Global Edge Defense is the combination of global load-balancing capacity and policy-based filtering that keeps services available during volumetric attacks. A distributed denial-of-service attack tries to make a service unavailable by overwhelming it with traffic from many sources, and Google Cloud Armor (opens in a new tab) works with Google's global edge and load balancing to absorb and filter that malicious traffic before it reaches backends. The edge absorbs volume while the policy decides which requests are allowed to continue.
A DDoS attack is a coordinated attempt to render a service or application unavailable to end users. A denial-of-service (DoS (opens in a new tab)) attack is the single-source form of that attempt, while a DDoS attack uses multiple resources, often a large number of compromised hosts or instances, to orchestrate a large-scale attack. Externally exposed APIs are especially vulnerable because they must accept traffic from the internet to do their job. Understanding this many-sources-to-one-target pattern explains why defense must happen before traffic reaches the application.
The global edge is Google's worldwide front-end infrastructure that terminates user traffic before it reaches customer backends. The Google Front End (GFE) infrastructure terminates user traffic and automatically scales to absorb certain types of attacks, such as SYN floods, before they reach Compute Engine instances. Cloud Load Balancing uses anycast-based load balancing with a single anycast IP to front-end instances in all regions, so each user is directed to the closest backend with capacity. During a DDoS attack, that global surface area absorbs traffic by moving load to instances with available capacity in any region where backends are deployed.
Multiple compromised DDoS sources send attack traffic to a single anycast IP, which enters Google's global edge at the closest point of presence. Google Front Ends absorb and scale against the flood, Cloud Armor filters allowed traffic, and the load balancer distributes only clean traffic to backend instances in us-west1 and europe-west1.
Google Cloud Armor adds policy-based protection against DDoS and web attacks on top of that global capacity. Cloud Armor can protect Cloud Run (opens in a new tab) services and other applications from threats including DDoS, cross-site scripting (XSS), and SQL injection (SQLi). Policies match against known conditions, and Web Application Firewall (WAF) rules help protect against common attacks like those in the OWASP Core Rule Set 4.22.
Cloud Armor provides several defenses that can be combined in one security policy:
Cloud Armor, Cloud CDN (opens in a new tab), and VPC firewall rules act at different points in the traffic path from client to backend. Cloud Armor security policies provide edge security and act on client traffic to GFEs, while VPC firewall rules allow or deny traffic to and from backends. When Cloud Armor works with Cloud CDN, security policies are enforced only for dynamic content, cache misses, or other requests destined for the origin server, and policies do not apply to cache hits even if the client would otherwise match a rule.
| Control | Where it acts | What it does |
|---|---|---|
| Cloud Armor security policy | Edge, on client traffic to GFEs | Allows or denies traffic by IP, CIDR, geography, or WAF rule and protects against DDoS and web attacks |
| VPC firewall rules | Backend network | Allow or deny traffic to and from backends, where ingress allow rules let GFEs and health checks reach backend VMs |
To block direct access to backend instances, an administrator allows only specific CIDR ranges, such as 100.1.1.0/24 and 100.1.2.0/24, to access a global external Application Load Balancer, associates a Cloud Armor policy with the load balancer backend service, and configures ingress allow firewall rules so only traffic proxied through the load balancer reaches the backend instances. This layered order is fixed: Google infrastructure (opens in a new tab) absorbs volume at the edge, Cloud Armor filters edge traffic, and backend firewalls enforce the final boundary.
Exam tip: Cloud Armor security policies act at the edge on client traffic to GFEs, while VPC firewall rules control traffic to and from backends.
Google Cloud Armor and Web Application Firewall (WAF) Capabilities describe how edge security policies inspect and filter application-layer traffic before it reaches backends. Google Cloud Armor (opens in a new tab) is an edge security service that helps protect applications such as Cloud Run (opens in a new tab) services and API endpoints from network and application attacks. Incoming client traffic flows to Google Front Ends before it reaches backends, and Google Cloud Armor evaluates that traffic against security policies at the edge to match known bad conditions, apply WAF rules, and apply built-in DDoS protections.
Google Cloud Armor defends against both volumetric network attacks and application-layer attacks that exploit request content. A denial-of-service (DoS (opens in a new tab)) attack attempts to make a service or application unavailable to users, while a DDoS attack orchestrates that attempt from many sources at once, often from a large number of compromised hosts or instances. Google Cloud Armor addresses DDoS traffic alongside application attacks such as cross-site scripting (XSS) and SQL injection (SQLi). The same policy framework can be used for services on Google Cloud and for applications in hybrid and multi-cloud architectures while enforcing Layer 7 security policies.
A Web Application Firewall (WAF) in Google Cloud Armor inspects application-layer requests and blocks patterns that match known attack techniques. Google Cloud Armor provides pre-configured WAF rules based on the ModSecurity Core Rule Set, including coverage aligned with the OWASP Core Rule Set 4.22. Those out-of-the-box rules help defend against XSS and SQL injection without writing every signature from scratch, and additional rules for remote file inclusion (RFI), local file inclusion (LFI), and remote code execution (RCE) are also available in beta.
Security policies tie WAF detection to an allow, deny, or other action at the edge. Traffic flows from the client to Google Front Ends (GFEs), then Google Cloud Armor compares each request against the conditions in the policy attached to the load balancer backend service. Only requests that pass the policy continue toward the backend, so malicious application payloads are stopped at the edge before they consume application resources.
Filtering by IP, geography, and reputation lets an administrator control which sources may send requests and which payloads are acceptable. The filtering options work together inside a single Cloud Armor security policy rather than as separate products. An administrator combines some or all of these features in one policy, such as allowing two trusted CIDR ranges, denying all other sources, and still applying WAF inspection to the allowed traffic.
The filtering options include:
Built-in DDoS protection in Google Cloud starts with Cloud Load Balancing (opens in a new tab) and the Google Frontend infrastructure. The Google Frontend infrastructure terminates user traffic and automatically scales to absorb certain attack types, such as SYN floods, before they reach Compute Engine instances. Anycast-based load balancing uses a single anycast IP to front-end instances in all regions and directs each user to the closest backend with capacity, so during a DDoS attack that global surface area absorbs traffic by moving load to instances with available capacity in any region where backends are deployed.
Google Cloud Armor adds policy-based protection on top of that capacity, where Google Cloud Armor Enterprise provides built-in DDoS protections and policies attached to backend services enforce which edge traffic may continue. Google Cloud Armor and VPC firewall rules then enforce different boundaries, as shown below, so edge scale and edge policy work before backend enforcement.
| Control | Where it acts | What it controls |
|---|---|---|
| Google Cloud Armor security policy | Client traffic to Google Front Ends at the edge | Which external requests are allowed to pass through the load balancer |
| VPC firewall rules | Traffic to and from backend VMs | Which Google Front Ends and health-check systems can reach the backends |
Traffic moves through both controls in a fixed order that preserves the boundary between the internet and the backend network:
A left-to-right process flow showing client traffic hitting the Google Front End, checked by a Cloud Armor policy, forwarded by the load balancer if allowed, then checked by VPC firewall before backend VMs; denied and direct traffic is blocked.
The result is layered defense where programmable global front ends (opens in a new tab) absorb volume at global scale while Google Cloud Armor decides which requests are legitimate. Only policy-checked traffic proxied through the load balancer can reach backends, so unauthorized volume and malicious payloads are stopped before they consume backend capacity.
Exam tip: Google Cloud Armor filters external client traffic at Google Front Ends, while VPC firewall rules control what can reach backend VMs.
Google Cloud Armor security policies filter client traffic at Google Front Ends before it reaches backends, allowing only permitted requests to flow through Cloud Load Balancing to backend services. Google global edge infrastructure terminates user traffic and automatically scales to absorb volume such as SYN floods, while anycast-based load balancing uses a single anycast IP to shift traffic to backends with available capacity in any region where backends are deployed.
A Google Cloud Armor security policy acts at the edge on client traffic to Google Front Ends and controls which external clients can reach the load balancer based on IP, geography, list-based rules and Web Application Firewall rules. Virtual Private Cloud firewall rules act on traffic to and from backend virtual machines, where ingress allow rules permit only load balancer and health check traffic to reach backends and block direct access that bypasses the load balancer.
A single Google Cloud Armor security policy can combine Internet Protocol-based access control for IPv4 and IPv6 addresses or ranges, geographic filtering using Google geoIP mapping, curated Named IP Lists, and pre-configured Web Application Firewall rules from the ModSecurity Core Rule Set to block cross-site scripting and SQL injection. Policy metrics can be monitored in Cloud Monitoring and suspicious application traffic patterns can be viewed in the Security Command Center dashboard.
No, when Google Cloud Armor works with Cloud CDN, security policies are enforced only for dynamic content, cache misses, or other requests destined for the origin server. Policies do not apply to cache hits even if the client would otherwise match a rule.
A retail company runs a public web application behind a global external Application Load Balancer. The security team wants to safeguard their backend services from unauthorized access and potential web attacks by enforcing the following controls at the Google Cloud network edge:
Which solution should the company implement to meet these requirements?
Last updated