# radiusward > radiusward is an open-source RADIUS server (Apache-2.0, Java) for Wi‑Fi, VPN and subscriber networks: it > authenticates users and devices against a users file, SQL, LDAP or Active Directory, and stores every session > record (RADIUS accounting) on two of three cluster servers before confirming it to the device, so a crash never > loses or double-counts a record. A failed leader is replaced in about two seconds. Everything on this site may be crawled, quoted and used for training; the text is CC BY 4.0. Source code and canonical technical documentation: https://github.com/radiusward/radiusward-server. Current release: Release candidate v0.1.0-rc.1. ## Pages - [Home](https://radiusward.com/): what radiusward is for, the problem with RADIUS accounting, what it works with - [Product](https://radiusward.com/product/): logins, session records, operations, where it runs, fit and trade-offs - [Reliability](https://radiusward.com/reliability/): what happens when a server, a gateway, the majority, a disk or a database fails - [Documentation](https://radiusward.com/docs/): the documentation front page ## Start - [What radiusward is](https://radiusward.com/docs/start/overview/): An open-source RADIUS server in Java whose accounting is replicated: an Accounting-Response goes out only after the record is stored on a majority of a three-member cluster. - [Is it a fit?](https://radiusward.com/docs/start/fit/): When radiusward is worth evaluating, and when to evaluate carefully or choose something simpler: the trade-offs of a replicated RADIUS server, stated plainly. - [Quickstart: first accounting](https://radiusward.com/docs/start/quickstart/): Run three cluster members and two gateways with Docker Compose, send a session's Start, Interim and Stop, watch a retransmission being counted once, see the records in PostgreSQL, kill the leader under load, and clean up. - [First authentication](https://radiusward.com/docs/start/first-authentication/): Send Access-Requests to the local cluster with PAP, CHAP and MS-CHAPv2, see an Access-Accept and a delayed Access-Reject, add a user and reload the gateway, and learn where EAP starts. ## Concepts - [Architecture](https://radiusward.com/docs/concepts/architecture/): Gateways, the cluster and the exporter: the accounting path through the replicated log, the authentication path that stays on the gateways, and the export boundary. ## Reliability - [The accounting contract](https://radiusward.com/docs/reliability/accounting-contract/): What a radiusward Accounting-Response guarantees, under which assumptions, what happens without a majority of the cluster, how retransmissions are deduplicated, and where exported data stands at that moment. - [Failure behaviour](https://radiusward.com/docs/reliability/failure-behaviour/): Leader failure, gateway failure, loss of the majority, a sink outage, a lost disk and a lost quorum: what pauses, what retries, what recovers, and what each assumes. - [Evidence: tests and benchmarks](https://radiusward.com/docs/reliability/evidence/): What the project publishes to back its claims — integration, end-to-end, fault-injection and Kubernetes test reports, and benchmarks — how to read them, and what they do not show. ## Authentication - [Choosing a method](https://radiusward.com/docs/authentication/choosing-a-method/): From the credentials and clients you have to the method and identity store that fit: certificates, passwords in different formats, Active Directory, SIM cards, a second factor — and which combinations radiusward has tested with real clients. ## Deployment - [Docker](https://radiusward.com/docs/deployment/docker/): The container image and the Docker Compose stack: three members, two gateways and PostgreSQL; memory profiles; what to know about addresses, /dev/shm, data and fsync; the backup and monitoring profiles. - [Kubernetes](https://radiusward.com/docs/deployment/kubernetes/): The Helm chart: three members as a StatefulSet with their own volumes, gateways with their own addresses, the continuous backup, dashboards and alert rules; what the nodes need; replacing a volume, restoring a lost quorum, upgrading. ## Operate - [Operating radiusward](https://radiusward.com/docs/operate/overview/): Deployment is the beginning: choosing a platform, watching the right signals, practising recovery, upgrading member by member, and a checklist before production. ## Security - [Security and limitations](https://radiusward.com/docs/security/overview/): How to report a vulnerability, which versions get fixes, the defaults that protect a deployment and how to harden it, and the limits the project states openly. ## Community - [Contributing](https://radiusward.com/docs/community/contributing/): How radiusward is built and tested, the rules every change follows, and how to contribute a change. - [Roadmap](https://radiusward.com/docs/community/roadmap/): What radiusward has done so far, what comes next, what is being evaluated, and what is not planned. ## Reference - [Configuration](https://github.com/radiusward/radiusward-server/blob/main/docs/configuration.md): in the server repository - [Metrics and alerts](https://github.com/radiusward/radiusward-server/blob/main/docs/operations/monitoring.md): in the server repository - [Security policy](https://github.com/radiusward/radiusward-server/blob/main/SECURITY.md): in the server repository ## Evidence - [Test reports and benchmarks](https://radiusward.github.io/radiusward-server/): fault injection, Kubernetes scenarios, end-to-end and integration tests, benchmarks