Crates.io | selium-log |
lib.rs | selium-log |
version | 0.1.0 |
source | src |
created_at | 2024-05-02 03:58:27.853468 |
updated_at | 2024-05-02 03:58:27.853468 |
description | selium-log is an open-source implementation of a log-based message queue. |
homepage | https://selium.com |
repository | https://github.com/seliumlabs/selium |
max_upload_size | |
id | 1227441 |
size | 108,782 |
Selium is an extremely developer friendly, composable messaging platform with zero build time configuration.
First, create a new Cargo project:
$ cargo new --bin hello-selium
$ cd hello-selium
$ cargo add futures
$ cargo add -F std selium
$ cargo add -F macros,rt tokio
Copy the following code into hello-world/src/main.rs
:
use futures::{SinkExt, StreamExt};
use selium::{prelude::*, std::codecs::StringCodec};
#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
let connection = selium::client()
.with_certificate_authority("certs/client/ca.der")? // your Selium cert authority
.with_cert_and_key(
"certs/client/localhost.der",
"certs/client/localhost.key.der",
)? // your client certificates
.connect("127.0.0.1:7001") // your Selium server's address
.await?;
let mut publisher = connection
.publisher("/some/topic") // choose a topic to group similar messages together
.with_encoder(StringCodec) // allows you to exchange string messages between clients
.open() // opens a new stream for sending data
.await?;
let mut subscriber = connection
.subscriber("/some/topic") // subscribe to the publisher's topic
.with_decoder(StringCodec) // use the same codec as the publisher
.open() // opens a new stream for receiving data
.await?;
// Send a message and close the publisher
publisher.send("Hello, world!".into()).await?;
publisher.finish().await?;
// Receive the message
if let Some(Ok(message)) = subscriber.next().await {
println!("Received message: {message}");
}
Ok(())
}
For the purpose of testing, developing Selium, or otherwise jumping straight into the action, you can use the selium-tools
CLI to generate a set of self-signed certificates for use with mTLS.
To do so, install the selium-tools
binary, and then run the gen-certs
command. By default, gen-certs
will output the certs to certs/client/*
and certs/server/*
. You can override these paths using the -s
and -c
arguments respectively.
$ cargo install selium-tools
$ cargo run --bin selium-tools gen-certs
Next, open a new terminal window and start a new Selium server, providing the certificates generated in the previous step.
$ cargo install selium-tools
$ cargo install selium-server
$ cargo run --bin selium-server -- \
--bind-addr=127.0.0.1:7001 \
--ca certs/server/ca.der \
--cert certs/server/localhost.der \
--key certs/server/localhost.key.der
Finally, in our original terminal window, run the client:
$ cargo run
Included in the repository is a benchmarks
binary containing end-to-end benchmarks for the publisher/subscriber clients.
These benchmarks measure the performance of both encoding/decoding message payloads on the client, as well the responsiveness of the Selium server.
To run the benchmarks with the default options, execute the following commands:
$ cd benchmarks
$ cargo run --release
This will run the benchmarks with default values provided for the benchmark configuration arguments, which should produce a summary similar to the following:
$ cargo run --release
Benchmark Results
---------------------
Number of Messages: 1,000,000
Number of Streams: 10
Message Size (Bytes): 32
| Duration | Total Transferred | Avg. Throughput | Avg. Latency |
| 1.3476 Secs | 30.52 MB | 22.65 MB/s | 1347.56 ns |
If the default configuration is not sufficient, execute the following command to see a list of benchmark arguments.
$ cargo run -- --help
Selium is a brokered messaging platform, meaning that it has a client and a server component. Check
out the client
and server
crates for more details.
We also have the user guide that includes all of this information and much more. Our Getting Started chapters will step you through the process of setting up a secure, working Selium platform in 5 minutes or less.
We'd love your help! If there's a feature you want, raise an issue first to avoid disappointment. While we're happy to merge contributions that are in line with our roadmap, your feature may not quite fit. Best to check first.