In a nutshell, what does Colossus do?

Colossus provides a lightweight framework for doing high-performance network I/O on the JVM. It handles things like setting up servers, managing connections, and handling many of the low-level tasks required by applications.

Should I use Colossus?

Colossus has been built primarily around the use case of creating stateless, low-latency backend services. While you can certainly build any kind of system you want using it, if you’re looking to build a web application or something that is more likely to be CPU bound than IO bound, then Colossus will probably not give you any major advantage over frameworks like Play! or Spray.

Colossus is ideal for I/O heavy applications that do a lot of “bit pushing”. Proxies, load balancers, databases, and data streams are all examples of applications Colossus is well-suited for.

How is Colossus different from Netty?

At a high level Colossus is similar to Netty. Both frameworks use a pool of worker threads, each with their own NIO event loop. Colossus is different in that it is tightly coupled to Akka, allowing near-seamless communication between event-loops and actors.

Every component you can attach to an event-loop is also capable of hooking into the worker actor’s receive method, which opens the doors to mixed sync/async code.

How is Colossus different from Akka I/O?

The biggest difference is your code sits much closer to the event loop than with Akka I/O, which requires all your connection handling code to reside in actors. Being able to keep user-level code in-thread can result in significantly less overhead for low-level operations. With Akka I/O, simply getting data from the wire to your code means sending it through several actors. With colossus, your code is literally only a couple function calls away from NIO, which at very high loads makes a major difference in performance.

Akka is generally very fast; the fact that Colossus is entirely built on Akka speaks for that. But in many cases the actor model is not the most ideal when it comes to performance, and the small overhead of sending actor messages can add up very quickly when millions of messages are in transit. Therefore Colossus is built on the premise than inter-actor communication is a non-trivial operation and should be avoided unless it’s actually necessary for concurrency.

Current State of Things

Colossus is currently in it’s “pre-1.0” phase, meaning things are working, but the framework in under very heavy development. We’re working very hard to make Colossus as clean and usable as possible without making significant performance sacrifices.

Release Workflow

Currently we’re following an accelerated release workflow. Currently there is no backwards compatibility between versions, but we have somewhat formalized the release schedule to help everyone manage upgrading. Releases are versioned with the scheme :


Right now the major version is fixed, minor versions introduce breaking changes, and build versions aim to be backwards compatible (but may have small breaking changes if there’s a pressing need). Eventually we’ll move to a more solid semantic versioning scheme where only major releases introduce breaking changes


Colossus is released under the Apache License version 2.0.

This documentation is released under the Creative Commons Attribution 3.0 Unported License.

The source code for this page can be found here.