This repository is part of the source code of Wire. You can find more information at wire.com or by contacting [email protected].
You can find the published source code at github.com/wireapp/wire.
For licensing information, see the attached LICENSE file and the list of third-party licenses at wire.com/legal/licenses/.
No license is granted to the Wire trademark and its associated logos, all of which will continue to be owned exclusively by Wire Swiss GmbH. Any use of the Wire trademark and/or its associated logos is expressly prohibited without the express prior written consent of Wire Swiss GmbH.
This repository contains the source code for the Wire server. It contains all libraries and services necessary to run Wire.
Documentation on how to self host your own Wire-Server is not yet available but is planned. Federation is on our long term roadmap.
See more in "Open sourcing Wire server code".
- Content of the repository
- Architecture Overview
- Development setup
- How to run
wire-server
with "fake" external dependencies - How to run
wire-server
with real AWS services - Roadmap
This repository contains the following source code:
- services
- nginz: Public API Reverse Proxy (Nginx with custom libzauth module)
- galley: Conversations and Teams
- brig: Accounts
- gundeck: Push Notification Hub
- cannon: WebSocket Push Notifications
- cargohold: Asset (image, file, ...) Storage
- proxy: 3rd Party API Integration
- tools
- api-simulations: Run automated smoke and load tests
- makedeb: Create Debian packages
- bonanza: Transform and forward log data
- libs: Shared libraries
It also contains
- build: Build scripts and Dockerfiles for some platforms
- deploy: (Work-in-progress) - how to run wire-server in an ephemeral, in-memory demo mode
- doc: Documentation
The following diagram gives a high-level outline of the (deployment) architecture of the components that make up a Wire Server as well as the main internal and external dependencies between components.
Communication between internal components is currently not guarded by dedicated authentication or encryption and is assumed to be confined to a private network.
There are two options:
This requires a range of dependencies that depend on your platform/OS, such as:
- Haskell & Rust compiler and package managers
- Some package dependencies (libsodium, openssl, protobuf, icu, geoip, snappy, cryptobox-c, ...) that depend on your platform/OS
See doc/Dependencies.md for details.
Once all dependencies are set up, the following should succeed:
# build all haskell services
make
# build one haskell service, e.g. brig:
cd services/brig && make
The default make target (fast
) compiles unoptimized (faster compilation time, slower binaries), which should be fine for development purposes. Use make install
to get optimized binaries.
For building nginz, see services/nginz/README.md
If you don't wish to build all docker images from scratch (e.g. the alpine-builder
takes a very long time), ready-built images can be downloaded from here.
If you wish to build your own docker images, you need docker version >= 17.05 and make
. Then,
make docker-services
will, eventually, have built a range of docker images. See the Makefile
s and Dockerfile
s, as well as build/alpine/README.md for details.
Integration tests require all of the haskell services (brig,galley,cannon,gundeck,proxy,cargohold) to be correctly configured and running, before being able to execute e.g. the brig-integration
binary. This requires most of the deployment dependencies as seen in the architecture diagram to also be available:
- Required internal dependencies:
- cassandra (with the correct schema)
- elasticsearch (with the correct schema)
- redis
- Required external dependencies are the following configured AWS services (or "fake" replacements providing the same API):
- SES
- SQS
- SNS
- S3
- Cloudfront
- DynamoDB
Setting up these real, but in-memory internal and "fake" external dependencies is done easiest using docker-compose
. Run the following in a separate terminal (it will block that terminal, C-c to shut all these docker images down again):
deploy/docker-ephemeral/run.sh
Then, to run all integration tests:
make integration
Or, alternatively, make
on the top-level directory (to produce all the service's binaries) followed by e.g cd services/brig && make integration
to run one service's integration tests only.
You can use $WIRE_STACK_OPTIONS
to pass arguments to stack through the Makefile
s. This is useful to e.g. pass arguments to tasty or temporarily disable -Werror
without the risk of accidentally committing anything, like this:
WIRE_STACK_OPTIONS='--ghc-options=-Wwarn --test-arguments="--quickcheck-tests=19919 --quickcheck-replay=651712"' make integration
Note that tasty supports passing arguments vie shell variables directly.
See this README
Documentation, configuration, and code for this is not fully ready yet (please do not open an issue to ask about this!). More information on how to run wire-server
will be available here in the near future.
As a brief overview, it requires setting up
- database clusters (cassandra, redis, elasticsearch)
- external dependencies
- Amazon account with access to
- SES
- SQS
- SNS
- S3
- Cloudfront
- DynamoDB
- Nexmo/Twilio accounts (if you want to send out SMSes)
- Giphy/Google/Spotify/Soundcloud API keys (if you want to support previews by proxying these services)
- TURN servers (if you want to support Voice/Video calls)
- Amazon account with access to
- production-ready configuration for all services
- additional infrastructure configuration (DNS, SSL certificates, metrics, logging, etc)
- Deployment options