Fastly Next-Gen WAF Architecture and Deployment Overview

Web Application Firewall (WAF)API securitySecurity

This datasheet provides detail into the highly performant, patented architecture of the Fastly Next-Gen WAF, as well as information on the wide array of deployment options available.

On this page

Unified web app and API security for any environment

Fastly offers the most flexibly deployed WAF on the market and can protect your apps and APIs wherever they are—in containers, on-premises, in the cloud, or at the edge—with one integrated solution. Gain comprehensive protection without sacrificing performance or requiring dedicated headcount: the Fastly Next-Gen WAF (powered by Signal Sciences) simply works out of the box and is so effective almost 90% of our customers run us in full blocking mode. 

The Fastly Next-Gen WAF provides the proactive protection modern apps require while integrating into your DevOps and security toolchains for unparalleled visibility. Our flexible architecture can advance your application security strategy by providing developers, operations, and security teams insight into where and how your web applications and APIs are attacked.

Architecture overview

Next-Gen WAF Architecture Overview diagram

The Fastly Next-Gen WAF is a hybrid software as a service (SaaS) solution with three main components. This patented approach, developed by Signal Sciences, allows us to easily scale and protect even the highest volume applications and APIs without impacting performance.


Lightweight agents you deploy on your existing infrastructure to perform detection and decisioning against requests quickly and accurately.

Agents consist of a small daemon process and are designed to handle extremely heavy loads while making highly-performant and accurate detections and decisions locally. The agent also collects metadata about the malicious requests it has processed and shares that metadata with the Cloud Engine. We protect some of the highest volume sites on the Internet, where tens of thousands of agents collectively process trillions of production requests without impacting app or API performance. Agents block attacks before they hit applications or APIs and provide visibility into not only requests that come in but also server responses and anomalies that show how the application is behaving.


Optional but powerful component that pairs with our agents to enforce high performance and reliability.

Modules run on virtually any web server (NGINX, Apache, IIS, and more) or application language (.NET, Java, Python, PHP, .nodeJS, and more). The module is just a few hundred lines of code to ensure both reliability and extreme performance. Its sole job is to pass requests through to the agent and receive and enforce decisions from the agent to allow the request through to the application or log/block it (depending on the mode set in the console).

Cloud Engine

Cloud-hosted analytics backend enriches the agent asynchronously with intelligence gathered from external and proprietary sources to make dynamic, application-specific detections.

The Cloud Engine collects and analyzes anonymized attack data and telemetry from the many thousands of software agents across our customer base. The output from the Cloud Engine is used by the agent locally to perform better detection and make more aggressive blocking decisions. The agent decisioning is enhanced by our Network Learning Exchange (NLX) which shares confirmed malicious IP sources within the management console, alerting you to suspicious actors before they are a threat to your applications and APIs. Other feeds include external lists of malicious IPs and customers’ custom IP lists, all of which provide additional request context that enriches the agent decisioning. This visibility and context is shared via our API and native integrations with the DevOps tools your team already uses, including Slack, PagerDuty, Jira, and more as well as security tools like Elastic and Palo Alto Networks Cortex XSOAR. Metrics and event reporting for your entire application footprint are also readily available via dashboards in a unified management console.

Deployment options

Native deployment options for data center, cloud, containers, and serverless.

Deployment Option #1: Cloud and container-native

Agent-module pair installs at your web server, API gateway, or at the app level within minutes. Our agent is infrastructure-agnostic which provides you the flexibility to deploy where you need it, without worrying about dependencies on underlying languages or frameworks.

Next-Gen WAF Cloud Container diagram
Deploying in Kubernetes and service mesh

New application tools and frameworks, such as Kubernetes, are quickly moving companies into a DevOps-focused world. Companies now release code faster than ever before and Fastly offers flexible deployment options to fit within your container strategy with three “layers” where you can install our WAF in Kubernetes and four methods for how you deploy. Additionally, our native integrations with Envoy Proxy and Istio service meshes mean Fastly provides visibility into both north-south (client-server) and east-west (service-to-service) requests.

Install Method

Layer 1: Ingress Controller

Layer 2: Mid-Tier Service

Layer 3: App Tier

Agent + module in same app container

Agent + module in different containers

Agent in reverse proxy mode in same container as app

Agent in reverse proxy in sidecar container

Fastly fully supports deployments for:
NGWAF logo row - Architecture

Deployment Option #2: Data center and legacy application

Customers needing protection for legacy applications or those deployed in data centers typically choose one of two deployment options: install the Fastly Next-Gen WAF to inspect traffic prior to web requests reaching the app or API endpoint, or install our agent in reverse proxy mode. For example, our module can be installed at the load balancer (A10 Networks, HAProxy, NGINX) or at the API gateway (Ambassador, Kong, Cloudentity). Our agent can be deployed in reverse proxy mode for customers with requirements that prevent installation at the load balancer or API gateway.


A10 Next-Gen WAF, powered by Fastly

For superior on-premise protection, A10 Networks and Fastly have partnered to offer customers the A10 Next-Gen WAF, powered by Fastly. The A10 Thunder ADC offers best-in-class application delivery and protection against a vast array of advanced attacks with exceptional accuracy, ensuring service availability while reducing complexity and TCO.

Deployment Option #3: At the edge

The Fastly Next-Gen WAF is available on the Fastly Edge Cloud Network, allowing customers to enforce security controls as part of Fastly delivery services. The edge cloud deployment option is seamlessly integrated with Fastly’s caching layer, Varnish.

This provides protection and acceleration closer to users and shields origin systems from abusive attack traffic while delivering world-class performance. Our edge deployment is ideal for customers who are unable to install software on existing infrastructure and for those who want to take advantage of the performance benefits of Fastly’s global content delivery network (CDN). This deployment option also offers additional features including Layer 3 and 4 always-on DDoS protection and TLS management.


Deployment Option #4: Cloud WAF

Cloud WAF empowers you to quickly and easily protect web applications, APIs, microservices, and serverless applications—without installing software on your infrastructure. Once deployed, a simple DNS change to point application traffic to Cloud WAF is all that’s needed to enable the visibility and protection of the Fastly Next-Gen WAF for your applications. All web requests are redirected to our cloud enforcement layer where bad requests are detected and blocked. All good, legitimate traffic is then forwarded to your application origin server. Cloud WAF is ideal for customers wanting to add an easy-to-manage WAF without making upstream changes to their CDN layer.


Protection that’s committed to data privacy

Many leading financial services firms, healthcare companies, and others with strict data privacy requirements all utilize Fastly’s next-gen WAF because of our strong architecture built for data privacy. All sensitive data is handled entirely within the customer environment and only sanitized and redacted portions of requests that are marked as attacks or anomalies are then sent to the Fastly Cloud Engine. 

Once the agent identifies a potential attack or anomaly in a request, a set of fully customizable redactions are applied locally and then the agent sends only the redacted individual parameter of the request which contains the attack payload, as well as a few other non-sensitive or benign portions of the request, such as client IP, user agent, URI, etc. Our backend only collects the response’s metadata e.g. response codes, sizes, and times. We provide customers the ability to fully customize redaction policies and fields as needed. For additional protection, Fastly automatically enforces redaction of common sensitive data types—such as passwords, keys, GUIDs, and any type of PII or PHI—before the request is sent to our backend. 

Betterment logo

“It works straight out of the box, scales automatically, and does a great job at providing visibility while securing the application.”

Anson Gomes

Lead Security Engineer

DevOps and security toolchain integrations

The best path to success for effective application and API protection is to provide the same baseline of security data to development, operations, and security teams in the tools they’re already using. Fastly works with the industry’s best tools and platforms to provide real-time alerting into your DevOps and security toolchains and to ensure it’s easy for your teams to leverage our production security telemetry within your organization’s current tools and processes for further investigation and analysis.

Out-of-the-box technology integrations help teams make or continue their transition to modern development models and architectures. Our single-click integrations include the most common development and operations alerting engines, chat-ops, project management, and incident tracking systems.

Technology and platform integrations

Run the Fastly Next-Gen WAF anywhere

Load Balancers

Load Balancers row logos

Web servers

NGWAF logo-row WebServers


NGWAF logo-row IAAS


NGWAF logo-row PAAS


NGWAF logo-row Containers

Config Management

NGWAF logo-row Config-Mgmt

Feed integrations & partners

Send and receive data from the Fastly Next-Gen WAF

DevOps toolchain

NGWAF logo-row DevOps-Toolchain


Fastly Next-Gen WAF datasheet

Learn how our WAF protects against web layer attacks and integrates with DevOps tools.

Blog Post
The WAF Efficacy Framework

The WAF efficacy framework helps you measure the effectiveness of your WAF.

Analyst Report
Gartner® Magic Quadrant™ for WAAP Report

Fastly is a Cloud WAAP Challenger. Compare vendors in this report.

Analyst Report
Gartner® Peer Insights “Voice of the Customer”: WAAP

Fastly is the only vendor to be named a Customers’ Choice for five consecutive years.

Ready to get started?

Get in touch.