SPIRE (the SPIFFE Runtime Environment) is a tool-chain for establishing trust between software systems across a wide variety of hosting platforms. Concretely, SPIRE exposes the SPIFFE Workload API, which can attest running software systems and issue SPIFFE IDs and SVIDs to them. This in turn allows two workloads to establish trust between each other, for example by establishing an mTLS connection or by signing and verifying a JWT token.
The SPIRE project is under active development. If you encounter any problems, please open an issue or drop by our Slack channel.
Pre-built releases can be found at https://github.com/spiffe/spire/releases. These releases contain both server and agent binaries plus the officially supported plugins.
Alternatively you can build SPIRE from source
Before trying out SPIRE, we recommend becoming familiar with its architecture and design goals.
To provide a minimal example of how SPIRE can be used, we are going to set up an SPIRE Server and SPIRE Agent. We will use them to issue identities to a workload identified by being run under a specified unix user ID.
For simplicity we will install both SPIRE Server and SPIRE Agent on the same machine. In an actual deployment, these would typically run on different machines.
Note: This getting started guide assumes that you are running Ubuntu 16.04.
Get the latest tarball from here and then extract it into /opt/spire with the next commands:
$ sudo tar zvxf {your_downloaded_tarball.tar.gz}
$ sudo cp -r spire-0.2/. /opt/spire/
# Replace "spire-0.2" with the version downloaded
$ sudo chmod -R 777 /opt/spire/
Add spire-server and spire-agent to our $PATH for convenience:
$ sudo ln -s /opt/spire/spire-server /usr/bin/spire-server
$ sudo ln -s /opt/spire/spire-agent /usr/bin/spire-agent
After putting the agent and server binaries at the proper location, we have to configure them. The SPIRE Server relies on plugins for much of its functionality. Plugin configurations can be found under the plugins { ... } section in /opt/spire/conf/server/server.conf.
For plugins which are not built in (see the reference guide for a list of built-ins), ensure that the corresponding configuration includes the path to the appropriate plugin binary:
plugin_cmd = "/path/to/plugin_binary"
Every SVID issued by a SPIRE installation is issued from a common trust root. SPIRE provides a pluggable mechanism for how this trust root can be retrieved. By default, it will use a key distributed on disk. The release includes a dummy CA key that we can use for testing purposes, but the plugin must be configured to find it.
In server.conf, identify the UpstreamCA "disk" { .. } plugin configuration section, and modify key_file_path and cert_file_path appropriately:
key_file_path = "/opt/spire/conf/server/dummy_upstream_ca.key"
cert_file_path = "/opt/spire/conf/server/dummy_upstream_ca.crt"
Prebuilt binaries may not have a default .data directory. You will need to create it.
$ mkdir -p /opt/spire/.data
The SPIRE Server reference guide covers in more detail the specific configuration options and plugins available.
When connecting to the SPIRE Server for the first time, the agent uses a configured X.509 CA certificate to verify the connection. SPIRE releases come with a "dummy" certificate in the client and server. For a production implementation, a separate key would be generated for the server and certificate to be bundled with the agent.
Edit /opt/spire/conf/agent/agent.conf so it looks for the trust bundle at the right path:
trust_bundle_path = "/opt/spire/conf/agent/dummy_root_ca.crt"
As with the server, SPIRE agent relies on plugins which can be configured at /opt/spire/conf/agent/agent.conf.
The SPIRE Agent reference guide covers in more detail the specific configuration options and plugins available.
We will start a server and join an agent to it using the join token attestation method. A join token is a manually generated single-use token that can be used to authenticate a connection. In more sophisticated implementations, SPIRE can be configured to use platform-specific mechanisms to authenticate an Agent to a Server.
Start your server.
$ spire-server run \
-config /opt/spire/conf/server/server.conf
In a different terminal generate a one time Join Token via spire-server token generate sub command. Use the -spiffeID option to associate the Join Token with spiffe://example.org/host SPIFFE ID.
$ spire-server token generate \
-spiffeID spiffe://example.org/host
# Token: aaaaaaaa-bbbb-cccc-dddd-111111111111
The Join Token will be used for node attestation and the associated SPIFFE ID will be used to generate the SVID of the attested node.
The default ttl of the Join Token is 600 seconds. We can overwrite the default value through -ttl option.
In the same terminal start the agent using the previously generated token so we can join it with the server.
$ spire-agent run \
-config /opt/spire/conf/agent/agent.conf \
-joinToken {your previously generated token}
We need to register the workload in the server so we can define under which attestation policy we are going to grant an identity to the workload. Since we are going to register it using an uid unix selector that will be mapped to a target SPIFFE ID, we first need to create a new user that we will call workload.
In a new terminal, create the user:
$ sudo useradd workload
Get the id so we can use it in the next step.
$ id -u workload
Create a new registration entry with spire-server entry create providing the workload user id.
$ spire-server entry create \
-parentID spiffe://example.org/host \
-spiffeID spiffe://example.org/host/workload \
-selector unix:uid:{workload user id from previous step}
At this point, the registration API has been called and the target workload has been registered with the SPIRE Server. We can now call the workload API using a command line program to request the workload SVID from the SPIRE Agent.
We will simulate the workload API interaction and retrieve the workload SVID bundle by running the api subcommand in the agent. Run the command as the user workload that we created in the previous step.
Note: If you are running on Vagrant you will need to run
sudo -ifirst.
$ su -c "spire-agent api fetch x509" workload
# SPIFFE ID: spiffe://example.org/host/workload
# SVID Valid After: yyyy-MM-dd hh:mm:ss +0000 UTC
# SVID Valid Until: yyyy-MM-dd hh:mm:ss +0000 UTC
# CA #1 Valid After: yyyy-MM-dd hh:mm:ss +0000 UTC
# CA #1 Valid Until: yyyy-MM-dd hh:mm:ss +0000 UTC
Optionally, you may write the SVID and key to disk with -write in order to examine them in detail with openssl.
$ su -c "spire-agent api fetch x509 -write /opt/spire/" workload
$ openssl x509 -in /opt/spire/svid.0.pem -text -noout
# Certificate:
# Data:
# Version: 3 (0x2)
# Serial Number: 4 (0x4)
# Signature Algorithm: sha256WithRSAEncryption
# Issuer: C=US, O=SPIFFE
# Validity
# Not Before: Dec 1 15:30:54 2017 GMT
# Not After : Dec 1 16:31:04 2017 GMT
# Subject: C=US, O=SPIRE
# Subject Public Key Info:
# Public Key Algorithm: id-ecPublicKey
# Public-Key: (521 bit)
# pub:
# 04:01:fd:33:24:81:65:b9:5d:7e:0b:3c:2d:11:06:
# aa:a4:32:89:20:bb:df:33:15:7d:33:55:13:13:cf:
# e2:39:c7:fa:ae:2d:ca:5c:d1:45:a1:0b:90:63:16:
# 6e:b8:aa:e9:21:36:30:af:95:32:35:52:fb:11:a5:
# 3a:f0:c0:72:8f:fa:63:01:95:ec:d9:99:17:8c:9d:
# ca:ff:c4:a7:20:62:8f:88:29:19:32:65:79:1c:b8:
# 88:5d:63:80:f2:42:65:4b:9e:26:d0:04:5a:58:98:
# a3:82:41:b0:ab:92:c9:38:71:00:50:c5:6d:3f:ab:
# 46:47:53:92:eb:be:42:55:44:1a:22:0b:ef
# ASN1 OID: secp521r1
# NIST CURVE: P-521
# X509v3 extensions:
# X509v3 Key Usage: critical
# Digital Signature, Key Encipherment, Key Agreement
# X509v3 Extended Key Usage:
# TLS Web Server Authentication, TLS Web Client Authentication
# X509v3 Basic Constraints: critical
# CA:FALSE
# X509v3 Subject Alternative Name:
# URI:spiffe://example.org/host/workload
# Signature Algorithm: sha256WithRSAEncryption
# 98:5e:33:14:ff:8e:77:40:1d:da:68:13:34:65:66:29:d0:f3:
# fa:c7:e5:45:58:4c:13:49:ad:47:4b:8e:ff:ad:e5:72:ca:7d:
# 45:ac:c8:88:3d:66:63:3f:f7:56:0e:34:df:9c:51:9f:7d:b9:
# 99:6f:a2:c8:78:bf:08:8c:02:17:ec:42:b8:5c:a9:e6:58:5a:
# cb:0f:16:3f:85:8a:08:20:2c:23:61:e3:89:48:f1:f0:bc:73:
# 2a:c0:9c:29:0e:ed:d8:2f:53:2c:82:67:70:6b:14:a1:eb:43:
# 1a:c5:04:0d:82:5b:f4:aa:3b:c5:37:db:22:17:97:ff:dc:d8:
# 01:27:44:29:18:1f:76:a3:9e:6a:50:31:5a:65:09:91:d7:8a:
# 79:03:0c:e9:22:f9:6c:15:02:db:a9:e2:fc:73:15:82:3a:0e:
# dd:4f:e5:04:b6:84:31:71:0d:ee:c5:b5:5a:21:d0:a9:8d:ec:
# 8c:4d:95:f2:43:b3:e9:ae:81:db:56:37:a2:74:23:69:05:1a:
# 2c:c8:11:09:40:18:67:6f:77:ff:57:ea:73:cd:49:9d:ba:6c:
# 85:70:d7:5c:a5:ba:46:0e:86:a2:c1:1d:27:f2:7a:2d:c1:4b:
# 16:87:b2:97:2f:98:ed:80:2a:5e:62:f4:7f:87:82:ff:67:96:
# e6:2e:fa:a1
SPIRE provides a beta implementation of the Envoy Secret Discovery Service (SDS). SDS can be used to transparently install and rotate TLS certificates and trust bundles in Envoy. Please see the SPIRE Agent configuration guide for more information.
If you have any questions about how SPIRE works, or how to get it up and running, the best place to ask questions is the SPIFFE Slack Organization. Most of the maintainers monitor the #spire channel there, and can help direct you to other channels if need be. Please feel free to drop by any time!
The SPIFFE community, and Scytale in particular, maintain the SPIRE project. Information on the various SIGs and relevant standards can be found in https://github.com/spiffe/spiffe.
The SPIFFE and SPIRE governance policies are detailed in GOVERNANCE
In case you're wondering where this project is headed, here is the SPIRE major feature roadmap:
JWT-SVID: Support in SPIRE for generating JWT-based SVIDs through the Workload API. Pending finalization of the JWT-SVID specification in SIG-Spec. Target date: end of July 2018.DoneHigh Availability: Support for clustering multiple SPIRE servers together within a single trust domain. Target date: end of October 2018.Done- Federation: Support for federating trust domains by joining SPIRE servers together. Pending finalization of the Federation API specification in SIG-Spec. Target date: end of
August 2018March 2019. Update: Experimental functionality available now based on latest draft spec. - Nested SPIRE: Support for chaining SPIRE server clusters together. Target date: end of February 2019.