Rooms

Rooms are best for session-based gameplay where the match between players takes place in a short-lived environment.

Use Case

A good example is a first person shooter multiplayer match. The match takes place between two teams in a single game session, and players enter through a lobby and matchmaking. When the match is concluded, the multiplayer environment the match took place in is closed and players return to a lobby.

This is one example of how Rooms can be used, but it is by no means the only use case. The important distinction between Rooms and Worlds (see below) is that Rooms are relatively short-lived and are meant to be created and closed by the game client through the coherence SDK.

See Rooms API.

Worlds

Worlds, as opposed to Rooms, are long-lived and permanent multiplayer environments provided by coherence. Using the Developer Portal, your project will easily define and manage your World configurations.

See Manage Worlds.

Use Case

A good example of a World is a permanent environment for an Massively Multiplayer Game (MMO). Regardless of the number of players connected, the environment is always available, and players can connect and disconnect at will.

Entities can be permanently saved in the world so that even if there are no active connections, they are persisted when players do connect.

See Worlds API.

Rooms and Worlds Work Together

Your project does not have to choose one-or-the-other. A project in coherence can contain both World and Rooms.

Use Case

A good example of this scenario is again, our MMO. Although players connect to a permanent and persistent World, they may enter a dungeon instance with other players. These dungeon instances can be Rooms.

The primary difference in the configuration and usage of Room and Worlds is that Worlds are managed in the Developer Portal, whereas Rooms are created and managed through the SDK.

coherence is currently in private preview. Some stability and performance issues may still be present and being ironed out. Additionally, more features are planned for the public release.

Current features

General

• Custom UDP transport layer using bit streams with reliability

• Smooth state replication

• Server-side, client-side, distributed authority

• Connected entity support

• Fast authority transfer

• Remote messaging (RPC)

• Persistence

• Multiple examples and showcases

• Verified support for Windows, macOS, Linux, Android, iOS and WebGL

• Support for rooms and worlds

SDK

• Unity SDK with intuitive no-code layer

• Per-field adjustable interpolation and extrapolation

• Input queues

• Easy deployment into the cloud

• SDK source included, no 3rd-party libraries

Optimization and performance

• Per-field compression and quantization

• Per-field sampling frequency adjustable at runtime

• Unlimited per-field levels of detail

• Areas of interest

• Accurate SimulationFrame tracking

Online

• Developer portal with server and service configurator

• Multiple regions (US East, EU Central)

• Player accounts

• Key-value store

• Matchmaking

• Ability to deploy one replication server and one simulation server per environment

• Prometheus and Grafana integration

Coming soon

SDK

• Multi-room simulators

• Input queue UX improvements

• Network Profiler

Online

• Additional regions

• Support for multiple simulators and replicators in a single project

• Dashboard with usage statistics

Mid-term roadmap

• Support for lean pure C# clients and simulators without Unity

• Peer-to-peer (without replication server) with NAT punch-through

• TCP fallback support

• WebSockets support

• MTU detection

• Packet replay

• Ability to deploy multiple simulation servers per environment

• Player analytics

• Developer portal graphs and analytics

• Simulator authentication

• Bare-metal and cloud support

Roadmap

• JavaScript SDK

• Unreal Engine SDK

• Multiple replication servers per game world

• Customer-specific serialization

• User-space load-balancing (SDK framework)

• Game world map with admin interface

• Anti-cheat functionality

• Advanced transaction logs (audit trail)

• Schema versioning (hot updates)

Get started

coherence is a network engine, platform and a series of tools to help anyone create a multiplayer game. Our mission is to give any game developer, regardless of how technical they are, the power to make a connected game.

Update an existing project

If you are an existing user and looking to update, you can find our change log.

Tutorial project

The Network Playground is a collection of scenes showing you how to use various features of the coherence Unity SDK. It shows you how to synchronize transforms, physics, persistence, animations, AI navigation and send network commands.

Existing or new Unity project

You can follow our step-by-step guide to learn how to install coherence in Unity, set up your scene, prefabs, interactions, as well as deploy your project to be shared with your friends.

Join the community

Join our community Discord for community chatter and support.

• Join our official Developer Discord channel.

• Contact us at devrel@coherence.io

Basic Terms

Replication Server ("Replicator")

A lean and performant server that keeps the state of the world and replicates it efficiently between various simulators and game clients. The Replicator usually runs in the coherence Cloud, but developers can start it locally from the command line or the Unity Editor.

Game Client

A build of the game. To connect to coherence, it will use the coherence SDK.

Simulation Server ("Simulator")

A version of the game client without the graphics ("headless client") optimized and configured to perform server-side simulation of the game world. When we say something is simulated on the server, we mean it is simulated on one or several simulators.

Schema

A text file defining the structure of the world from the network's point of view. The schema is shared between the replicators, simulators and game clients. The world is generally divided in components and archetypes.

**** Code Generation

The process of generating code specific to the game engine that takes care of network synchronization and other network-specific code. This is done using a CLI tool called _Protocol Code Generator _ that takes the schema file and generates code for various engines (e.g. C# for Unity).

State Replication

The process of making sure the state of the world is eventually the same on the replicator, simulators and game clients, depending on their areas of interest.

State Replication

coherence works by sharing game world data via a Replication Server in the cloud and passing it to the connected clients.

The clients and simulators can define areas of interest (LiveQueries), levels of detail, varying simulation and replication frequencies and other optimization techniques to control how much bandwidth and CPU power is used in different situations.

The game world can be run using multiple simulators that split up simulation functions or areas of the world accordingly.

The platform handles scaling, synchronization, persistence and load balancing automatically.

Overview

coherence is a network engine, platform and a series of tools to help anyone create a multiplayer game.

• Fast network engine with cloud scaling, state replication, persistence and auto load balancing.

• Easy to develop, iterate and operate connected games and experiences.

• SDK allows developers to make multiplayer games using Windows, Linux or Mac, targeting desktop, console, mobile, VR or the web.

• Game engine plugins and visual tools will help even non-coders create and quickly iterate on a connected game idea.

• Scalable from small games to large virtual worlds running on hundreds of servers.

• Game-service features like user account, key-value stores and matchmaking.

Network Engine

At the core of coherence lies a fast network engine based on bitstreams and a data-oriented architecture, with numerous optimization techniques like delta compression, quantization and ("Level of Detail") to minimize bandwidth and maximize performance.

Authority Models

The network engine supports multiple authority models:

• Client authority

• Server authority

• Server authority with client prediction

• Authority handover (request, steal)

• Distributed authority (multiple simulators with seamless transition)

• Deterministic client prediction with rollback (""): coming in a future release

Persistence

coherence supports persistence out of the box.

This means that the state of the world is preserved no matter if clients or simulators are connected to it or not. This way, you can create shared worlds where visitors have a lasting impact.

Engine Support

Game Engine Support

coherence currently supports Unity. For custom engine integration,. For updates regarding Unreal Engine support, please check the page.

Unity Requirements

• Unity 2020.3.0f1 or later.

• A Windows, Linux or macOS system.

Install Video Tutorial

Unity Package Manager

1. Add a scoped registry

First, open Project Settings.

Under Package Manager, add a new Scoped Registry with the following fields:

• Name: coherence

• URL: https://registry.npmjs.org

• Scope(s): io.coherence.sdk

• Enable Preview / Pre-release Packages: checked

• Show Dependencies: checked

Click Apply.

2. Find and install the coherence package

Now open the Package Manager.

Click Packages and My Registries.

Under coherence, click Install.

Alternative Method: Edit manifest.json manually

If you want to install coherence manually, go to the folder of your project and open the file /Packages/manifest.json.

Copy paste the lines surrounded by comments that look like this:/* comment */.

{
  "dependencies": {
    "com.unity.collab-proxy": "1.3.9",
    "com.unity.ide.rider": "2.0.7",
    "com.unity.ide.visualstudio": "2.0.7",
    "com.unity.ide.vscode": "1.2.3",
    "com.unity.test-framework": "1.1.24",
    "com.unity.textmeshpro": "3.0.1",
    "com.unity.timeline": "1.4.6",
    "com.unity.ugui": "1.0.0",

    /*** ADD THIS START ***/
    "io.coherence.sdk": "0.7.0",
    /*** ADD THIS END ***/
        
    "com.unity.modules.ai": "1.0.0",
    "com.unity.modules.androidjni": "1.0.0",
    "com.unity.modules.animation": "1.0.0",
    "com.unity.modules.assetbundle": "1.0.0",
    "com.unity.modules.audio": "1.0.0",
    "com.unity.modules.cloth": "1.0.0",
    "com.unity.modules.director": "1.0.0",
    "com.unity.modules.imageconversion": "1.0.0",
    "com.unity.modules.imgui": "1.0.0",
    "com.unity.modules.jsonserialize": "1.0.0",
    "com.unity.modules.particlesystem": "1.0.0",
    "com.unity.modules.physics": "1.0.0",
    "com.unity.modules.physics2d": "1.0.0",
    "com.unity.modules.screencapture": "1.0.0",
    "com.unity.modules.terrain": "1.0.0",
    "com.unity.modules.terrainphysics": "1.0.0",
    "com.unity.modules.tilemap": "1.0.0",
    "com.unity.modules.ui": "1.0.0",
    "com.unity.modules.uielements": "1.0.0",
    "com.unity.modules.umbra": "1.0.0",
    "com.unity.modules.unityanalytics": "1.0.0",
    "com.unity.modules.unitywebrequest": "1.0.0",
    "com.unity.modules.unitywebrequestassetbundle": "1.0.0",
    "com.unity.modules.unitywebrequestaudio": "1.0.0",
    "com.unity.modules.unitywebrequesttexture": "1.0.0",
    "com.unity.modules.unitywebrequestwww": "1.0.0",
    "com.unity.modules.vehicles": "1.0.0",
    "com.unity.modules.video": "1.0.0",
    "com.unity.modules.vr": "1.0.0",
    "com.unity.modules.wind": "1.0.0",
    "com.unity.modules.xr": "1.0.0"
  }, /* add this comma if not already present */
  
  /*** ADD THIS SECTION START ***/
  "scopedRegistries": [
    {
      "name": "coherence",
      "url": "https://registry.npmjs.org",
      "scopes": [
        "io.coherence.sdk"
      ]
    }
  ]
  /*** ADD THIS SECTION END ***/
  
}

When you install the coherence Unity Package, all the services for the SDK are installed in the background as well.

You will then see this package in the Package Manager under "My Registries".

The coherence SDK has some dependencies on other Unity packages you can see in the image above. If you're already using these in your project, you might have to adjust their version number (Unity will tell you about this).

When you successfully install the coherence SDK you'll get this quickstart winodow pop-up and you'll be good to go.

Now we can build the project and try out network replication locally.

This example will show you how to launch a local replication server and connect multiple instances.

Running a Replication Server Locally

You can run a local replication server from the coherence menu by clicking:

coherence -> Server -> Run Local Worlds Server.

This will open a new terminal window with the replication server and world created in it.

Build and Connect

Now it's time to make a standalone build and test network replication.

Open the Build Settings window (File -> Build Settings). Click on Add Open Scenes to add the current scene to the build. Click Build and Run.

Select a folder (e.g. builds) and click OK.

When the build is done, start another instance of the executable (or run the project in the Game Window in Unity).

Click Connect on both clients. Now try focusing one and using WSAD keys. You will see the box move on the other side as well.

Connect across the local network

If you want to connect to the local replication server from another local device (such as another PC, Mac, Mobile or VR device), you can find your IPv4 address and use that as your server address in the Connect dialog. These devices need to be connected to the same network.

You can find your IPv4 address by going to your command line tool and type ipconfig . Remember to include the port number, for example 192.168.1.185:32001.

Setup Video Tutorial

It's quick and easy to set up a networked scene from scratch using the coherence SDK. This example will show you the basic steps to sync up some moving characters.

Add these components to your scene to prepare it for network synchronization.

1. Add MonoBridge

coherence -> Scene Setup -> Create MonoBridge

This object takes care of connected GameObject lifetimes and allows us to develop using traditional MonoBehaviour scripts.

2. Add LiveQuery - Hierarchy

coherence -> Scene Setup -> Create LiveQuery

Creates a LiveQuery which queries the area around the local player to get required information from the replication server. You can surround your entire scene in one query or can attach it to an object such as the player or a camera.

3. Add the Sample UI

coherence -> Scene Setup -> Add Sample UI

Creates a Canvas (and Event System if not already present in the scene) with a sample UI that helps you connect to a local or remote replication server. You can create your own connection dialog, this one is just a quick way to get started.

Baking Video Tutorial

Overview

Out of the box, coherence will use C# Reflection to sync all the data at runtime. This is a great way to get started but it is very costly performance-wise.

coherence offers an automatic way of doing that called baking.

Bake schemas

Click on coherence -> Schema and Baking -> Bake Schemas.

This will go through all CoherenceSync components in the project and generate a schema file based on the selected variables, commands and other settings. It will also take into account any LODs added.

For every prefab with a CoherenceSync object, the baking process will generate a bespoke C# file in the coherence/baked folder in the project.

Adding that file to the prefab will make that prefab use bespoke generated code instead of C# reflection.

Activate baked schema on prefab

Once the Schema has been baked, you will be able to switch to baked mode in the CoherenceSync inspector.

The name of the baked script will be CoherenceSync[prefabName].

Modifying bound data, safe mode and the watchdog

When you bind to your script's fields and bake, coherence generates specific code that accesses your code directly, without using reflection. This means, whenever you change your scripts, you might break compilation.

For example, if you have a Health.cs script which exposes a public float health; field, and you mark health as a binding in the Bindings window, the baked script will access your component via type name, and your field via field name.

Your baked script might now reference your component:

This means that if you decide to change your component name (Health) or any of your bound field names (health), Unity script recompilation will fail. In this example, we will be deprecating health and adding health2 in its place.

Our watchdog is able to understand when this happens, and offer you a solution right away.

It will suggest you to bake in safe mode, and then diagnose the state of your prefabs. After a few seconds of script recompilation, you'll be presented with the diagnosis window.

In this window, you can easily spot bindings in your prefabs that are no longer valid. In our example, health is no longer valid since we've moved it elsewhere (or deleted it).

Click on the hand pointing button to open the bindings window, and take a look at your script:

Now, we can manually rebind our data: unbind health and bind health2. Once we do, we can now safely bake again.

In this section, we will learn how to prepare a prefab for network replication.

1. Create a prefab and move it to Resources

Add an asset and create a prefab of it. Make sure the prefab is in a Resources folder in your Unity project.

Here is an example:

1.1 Create a simple Cube in the scene

GameObject -> 3D Object -> Cube

1.2 Convert the Cube into a prefab

Create a Resources folder in your Project. Drag the Cube into the Resources folder to turn it into a prefab.

2. Add CoherenceSync to the prefab

The CoherenceSync component will help you prepare an object for network synchronization during design time. It also exposes an API that allows us to manipulate the object during runtime.

CoherenceSync will query all public variables and methods on any of the attached components, for example Unity components such as Transform, Animator , etc.. This will include any custom scripts such as PlayerInput and even scripts that came with Asset Store packages you may have downloaded.

3. Select variables to replicate

Select which variables you would like to sync across the network. Initially, this will probably be the Transform settings; position, rotation, scale.

Under Configure, click Select fields and methods.

In the Configuration dialog, select position, rotation and scale.

Close the Configuration dialog_._

4. Add an input script

This simple input script will use WASD or the Arrow keys to move the prefab around the scene.

Click on Assets -> Create -> C# Script.

Name it Move.cs. Copy-paste the following content into the file.

Wait for Unity to compile the file, then add it onto the prefab.

5. Disable Input on replicated object

We have added a Move script to the prefab. This means that if we just run the scene, we will be able to use the keyboard to move the object around.

But what happens on another client where this object is not authoritative, but rather replicated? We will want the position to be replicated over the network, without the keyboard input interfering with it.

Open the Events section in CoherenceSync. Add a new On Network Instantiation handler by clicking on the plus sign next to it.

Pull the Cube prefab into the Runtime Only / None (Object) field.

Now click the dropdown No Function and select Move -> bool enabled.

Leave the Boolean field unchecked.

6. Additional CoherenceSync settings

From the CoherenceSync component you can configure settings for Lifetime (Session-based or Persistent, Authority transfer (Request or Steal), Simulation model (Client Side, Server Side or Server Side with Client Input) and Adoption settings for when local persistent entities are orphaned.

• On Before Networked Instantiation (before the network is instantiated)

• On Networked Instantiation (when the GameObject is instantiated)

• On Networked Destruction (when the GameObject is destroyed)

• On Authority Gained (when authority over the GameObject is transferred to the local client)

• On Authority Lost (when authority over the GameObject is transferred to another client)

• On After Authority Transfer Rejected (when GameObject's Authority transfer was requested and denied).

• On Input Simulator Connected (when client with simulator is ready for Server-side with Client Input)

• On Input Owner Assigned (when InputOwner was changed is ready)

  Extras

7. Constraints

There are some constraints when setting up a prefab with CoherenceSync, hereafter referred to as Sync Prefab.

1. A Sync Prefab has one, and only one CoherenceSync component in its hierarchy

2. The CoherenceSync component must be at the Sync Prefab root

3. A Sync Prefab cannot contain instances of other Sync Prefabs

4. A hierarchy in a scene can contain multiple Sync Prefabs. However, such a hierarchy cannot be saved as a Sync Prefab as that would break rule 1-3.

7.1. Examples of valid setup

7.2. Examples of disallowed setup

Networked entities can be simulated either on a game client ("client authority") or a simulation server ("server authority).

Client Authority

Client authority is the easiest to set up initially, but it has some drawbacks:

• Higher latency. Because both clients have a non-zero ping to the replication server, the minimum latency for data replication and commands is the combined ping (client 1 to replications server and replication server to client 2).

• Higher exposure to cheating. Because we trust game clients to simulate their own entities, there is a risk that one such client is tampered with and sends out unrealistic data.

In many cases, especially when not working on a competitive PvP game, these are not really issues and are a perfectly fine choice for the game developer.

Client authority does have a few advantages:

• Easier to set up. No client vs. server logic separation in the code, no building and uploading of simulation servers, everything just works out of the box.

• Cheaper. Depending on how optimized the simulator code is, running a simulator in the cloud will in most cases incur more costs than just running a replication server (which is comparatively very lean).

Server Authority

Having one or several simulators taking care of the important world simulation tasks (like AI, player character state, score, health, etc.) is always a good idea for competitive PvP games.

Running a simulator in the cloud next to the replicator (the ping between them being negligible) will also result in lower latency.

The player character can also be simulated on the server, with the client locally predicting its state based on inputs. You can read more about how to achieve that in the section .

Remote Communication

coherence allows you to upload and share the builds of your games to your team, friends or adoring fans via an eay access play link.

Right now we support desktop (PC, Mac, Linux) and also WebGL, where you can host and instantly play your multiplayer game and share it around the world.

Build the game

Build your game to a local folder on your desktop as you would normally.

In the coherence menu in Unity select "Share -> Build Upload"

Upload to coherence

in this window you can select which platform the build is for an you also need to browse to the local path folder.

Click "Upload" or "Begin Upload" and coherence will confirm that it's okay to compress and upload the build to the webdashboard.

Share your build

Now that build has been updated (signified by the green tick), you can share it by enabing and share the public URL. Anyone with this link can access the build.

WebGL

If you uploaded a WebGL build then you can play it instantly from that public link for instant play.

Overview

The state of network entities that are currently not simulated locally (either because they are being simulated on another game client or on a simulator) cannot be affected directly.

Network can help us affect state indirectly, but for anything more involved, an authority transfer might be necessary.

Types of Authority Transfer

In the design phase, CoherenceSync objects can be configured to handle authority transfer in different ways:

• Request. Authority transfer may be requested, but it may be rejected by the receiving party.

• Steal. Authority will always be given to the requesting party on a FCFS ("first come first serve") basis.

• Not transferable. Authority cannot be transferred.

Note also that you need to set up Auto-adopt Orphan if you wan't orphans to be adopted automatically by nearest player.

When using Request, an optional callback OnAuthorityRequestedByConnection can be set on the CoherenceSync behaviour. If the callback is set, then the results of the callback will override the Approve Requests setting in the behaviour.

The request can be approved or rejected in the callback.

Requesting Authority in Code

Requesting authority is very straight-forward.

As the transfer is asynchronous, we have to subscribe to one or more Unity Events in CoherenceSync to learn the result.

The request will first go to the replication server and be passed onto the receiving simulator or game client, so it may take a few frames to get a response.

These events are also exposed to the Unity inspector in the Events on authority transfer section of the CoherenceSync behaviour.

Now we can finally deploy our schema and replication server into the cloud.

Upload schemas

In the Project Settings tab for coherence, click on Upload.

The status should change from "Unknown" to "In Sync".

Your project schema is now deployed with the correct version of the replication server already running in the cloud. You will be able to see this in your cloud dashboard status.

Run project

You can now build the project again and send the build to your friends for testing.

You will be able to play over the internet without worrying about firewalls and local network connections.

Cloud Deploy Video Tutorial

Now that we have tested our project locally, it's time to upload it to the cloud and share it with our friends and colleagues. To be able to do that, we need to create a free account with coherence.

Sign up or log in

Create an account or log into an existing one.

Create an organization

Create an organization.

Create project

Under the organization, create a new project.

Get portal token

Copy the token to your clipboard.

Copy Portal token into Unity

Open Unity and navigate to the Project Settings. Open the coherence tab_._

Paste the token into the Portal Token section.

Fetch the Runtime key

Once you have pasted the portal token successfully, you need to fetch the runtime token as well.

It is often useful to know when a synced property has changed its value. It can be easily achieved using the OnValueSyncedAttribute. This attribute lets you define a method that will be called each time a value of a synced member (field or property) changes in the non-simulated version of an entity.

Usage

Let's start with a simple example:

using Coherence.Toolkit;
using UnityEngine;
using UnityEngine.UI;

public class Player : MonoBehaviour
{
    [OnValueSynced(nameof(UpdateHealthLabel))]
    public float Health;

    public Text HealthLabel;

    public void UpdateHealthLabel(float oldHealth, float newHealth)
    {
        HealthLabel.text = newHealth.ToString();
        Debug.Log($"Player HP changed by: {newHealth - oldHealth}");
    }
}

Whenever the value of the Health field gets updated (synced with its simulated version) the UpdateHealthLabel will be called automatically, changing the health label text and printing log with a health difference.

This comes in handy in projects that use authoritative simulators. The client code can easily react to changes in the Player entity state introduced by the simulator, updating the visual representation (which the simulator doesn't need).

Limitations

The OnValueSynced feature can be used only on members of user-defined types, that is, there's no way to be notified about a change in the value of a Unity type member, like transform.position. This might however change in the future, so stay tuned!

Persistent objects are stored

All persistent objects remain in the world for the entire lifetime of the replication server and, periodically, the replication server records the state of the world and saves it to physical storage. If the replication server is restarted, then the saved persistent objects are reloaded when the replication server resumes.

Persistent vs. session-based entities

When we connect to a game world with a game client, the traditional approach is that all entities originating on our client are session-based. This means that when the client disconnects, they will disappear from the network world for all players.

A persistent object, however, will remain on the replication server even when the client or simulator that created or last simulated it, is gone.

This allows us to create a living world where player actions leave lasting effects.

In a virtual world, examples of persistent objects are:

• A door anyone can open, close or lock

• User-generated or user-configured objects left in the world to be found by others

• Game progress objects (e.g. in PvE games)

• Voice or video messages left by users

• NPC's wandering around the world using an AI logic

• Player characters on "auto pilot" that continue affecting the world when the player is offline

• And many, many more

A persistent object with no simulator is called an orphan. Orphans can be configured to be auto-adopted by clients or simulators on a FCFS basis.

Overview

coherence input queues are backed by a rolling buffer of inputs transmitted between the clients. This buffer can be used to build a fully deterministic simulation with a client side-prediction, rollback, and input delay. This game networking model is often called the GGPO (Good Game Peace Out).

Input delay allows for a smooth, synchronized netplay with almost no negative effect on the user experience. The way it works is input is scheduled to be processed X frames in the future. Consider a fighting game scenario with two players. At frame 10 Player A presses a kick button that is scheduled to be executed at frame 13. This input is immediately sent to Player B. With a decent internet connection, there's a big chance that Player B will receive that input even before his frame 13. Thanks to this the simulation is always in sync and can progress steadily.

Prediction is used to run the simulation forward even in the absence of inputs from other players. Consider the scenario from the previous paragraph - what if Player B doesn't receive the input on time? The answer is very simple - we just assume that the input state hasn't changed and progress with the simulation. As it turns out this assumption is valid most of the time.

Rollback is used to correct the simulation when our predictions turn out wrong. The game keeps historical states of the game for past frames. When an input is received for a past simulation frame the system checks whether it matches the input prediction made at that frame. If it does we don't have to do anything (the simulation is correct up to that point). If it doesn't match, however, we need to restore the simulation state to the last known valid state (last frame which was processed with non-predicted inputs). After restoring the state we re-simulate all frames up to the current one, using the fresh inputs.

Determinism

In a deterministic simulation, given the same set of inputs and a state we are guaranteed to receive the same output. In other words, the simulation is always predictable. Deterministic simulation is a key part of the GGPO model, as well as a lockstep model because it lets us run exactly the same simulation on multiple clients without a need for synchronizing big and complex states.

Implementing a deterministic simulation is a non-trivial task. Even the smallest divergence in simulation can lead to a completely different game outcome. This is usually called a desync. Here's a list of common determinism pitfalls that have to be avoided:

• Using Update to run the simulation (every player might run at a different frame rate)

• Using coroutines, asynchronous code, or system time in a way that affects the simulation (anything time-sensitive is almost guaranteed to be non-deterministic)

• Using Unity physics (it is non-deterministic)

• Using random numbers generator without prior seed synchronization

• Non-symmetrical processing (e.g. processing players by their spawn order which might be different for everyone)

• Relying on floating point numbers across different platforms, compilations or processor types

Quickstart guide

We'll create a simple, deterministic simulation using provided utility components.

Our simulation will synchronize the movement of multiple clients, using the rollback and prediction in order to cover for the latency.

Player implementation

Start by creating a Player component and a prefab for it. We'll use the client connection system to make our Player represent a session participant and automatically spawn the selected prefab for each player that connects to the server. The Player will also be responsible for handling inputs using the CoherenceInput component.

Our Player code looks as follows:

using Coherence.Toolkit;
using UnityEngine;

[RequireComponent(typeof(CoherenceSync))]
[RequireComponent(typeof(CoherenceInput))]
public class Player : MonoBehaviour
{
    // Movement speed in units/sec
    public float Speed = 5f;
    
    private CoherenceInput input;

    private void Awake()
    {
        input = GetComponent<CoherenceInput>();
    }

    // Retrieves the "movement" input state for a given frame
    public Vector2 GetMovement(long frame)
    {
        return input.GetAxisState("Mov", frame);
    }

    // Sets the "movement" state for the current frame
    public void SetMovement()
    {
        Vector2 movement = new Vector2(Input.GetAxis("Horizontal"), Input.GetAxis("Vertical")).normalized;
        input.SetAxisState("Mov", movement);
    }
}

The GetMovement and SetMovement will be called by our "central" simulation code. Now that we have our Player defined let's prepare a prefab for it. Create a game object and attach the Player component to it. Using the CoherenceSync inspector create a prefab. The inspector view for our prefab should look as follows:

A couple of things to note:

• A Mov axis has been added to the CoherenceInput which will let us sync the movement input state

• Unlike in the server-side input queues our simulation uses client-to-client communication, meaning each client is responsible for its entity and sending inputs to other clients. To ensure such behavior set the CoherenceSync > Simulation and Interpolation > Simulation Type to Client Side

• In the deterministic simulation, it is our code that is responsible for producing deterministic output on all clients. This means that the automatic transform position syncing is no longer desirable. To turn it off check the CoherenceSync > Manual Position Update

• In order for inputs to be processed in a deterministic way, we need to use the fixed simulation frames. Tick the CoherenceInput > Use Fixed Simulation Frames checkbox

• Make sure to use the baked mode (CoherenceInput > Use Baked Script) - inputs do not work in the reflection mode

Since our player is the base of the client connection we must set it as the connection prefab in the CoherenceMonoBridge and enable the global query:

State implementation

Before we move on to the simulation, we need to define our simulation state which is a key part of the rollback system. The simulation state should contain all the information required to "rewind" the simulation in time. For example, in a fighting game that would be a position of all players, their health, and perhaps a combo gauge level. In a shooting game, this could be player positions, their health, ammo, and a map objective progression.

In the example we're building player position is the only state. We need to store it for every player:

using UnityEngine;

public struct SimulationState
{
    public Vector3[] PlayerPositions;
}

Simulation implementation

Simulation code is where all the logic should happen, including applying inputs and moving our Players:

using Coherence.Toolkit;
using UnityEngine;

public class Simulation : CoherenceInputSimulation<SimulationState>
{
    protected override void SetInputs(CoherenceClientConnection client)
    {
        var player = client.GameObject.GetComponent<Player>();
        player.SetMovement();
    }

    protected override void Simulate(long simulationFrame)
    {
        foreach (CoherenceClientConnection client in AllClients)
        {
            var player = client.GameObject.GetComponent<Player>();
            var movement = (Vector3)player.GetMovement(simulationFrame);
            player.transform.position += movement * player.Speed * FixedTimeStep;
        }
    }

    protected override void Rollback(long toFrame, SimulationState state)
    {
        for (var i = 0; i < AllClients.Count; i++)
        {
            Transform playerTransform = AllClients[i].GameObject.transform;
            playerTransform.position = state.PlayerPositions[i];
        }
    }

    protected override SimulationState CreateState()
    {
        var simulationState = new SimulationState { PlayerPositions = new Vector3[AllClients.Count] };
        for (var i = 0; i < AllClients.Count; i++)
        {
            Transform playerTransform = AllClients[i].GameObject.transform;
            simulationState.PlayerPositions[i] = playerTransform.position;
        }

        return simulationState;
    }

    protected override void OnClientJoined(CoherenceClientConnection client)
    {
        SimulationEnabled = AllClients.Count >= 2;
        if (SimulationEnabled)
        {
            // Lets us rejoin the same simulation without restarting the app.
            StateStore.Clear();
        }
    }

    protected override void OnClientLeft(CoherenceClientConnection client)
    {
        SimulationEnabled = AllClients.Count >= 2;
    }
}

• SetInputs is called by the system when it's time for our local Player to update its input state using the CoherenceInput

• Simulate is called when it's time to simulate a given frame. It is also called during frame re-simulation after misprediction - don't worry though, the complex part is handled by the CoherenceInputSimulation internals - all you need to do in this method is apply inputs from the CoherenceInput to run the simulation

• Rollback is where we need to set the simulation state back to how it was at a given frame. The state is already provided in the state parameter, we just need to apply it

• CreateState is where we create a snapshot of our simulation so it can be used later in case of rollback

• OnClientJoined and OnClientLeft are optional callbacks. We use them here to start and stop the simulation depending on the number of clients

As a final step attach the Simulation script to the MonoBridge object on scene and link the MonoBridge back to the Simulation:

That's it! Once you build a client executable you can verify that the simulation works by connecting two clients to the replication server. Move one of the clients using arrow keys while observing the movement being synced on the other one.

Input in fixed network update

Due to the FixedNetworkUpdate running at different (usually lower) rate than Unity's Update loop, polling inputs using the functions like Input.GetKeyDown is susceptible to a input loss, i.e. keys that were pressed during the Update loop might not show up as pressed in the FixedNetworkUpdate.

To illustrate why this happens consider the following scenario: given Update running five times for each network FixedNetworkUpdate, if we polled inputs from the FixedNetworkUpdate there's a chance that an input was fully processed within the five Updates in-between FixedNetworkUpdates, i.e. a key was "down" on the first Update, "pressed" on the second, and "up" on a third one.

To prevent this issue from occurring you can use the FixedUpdateInput class:

[RequireComponent(typeof(CoherenceSync))]
[RequireComponent(typeof(CoherenceInput))]
public class Player: MonoBehaviour
{
    private FixedUpdateInput fixedUpdateInput;
    private CoherenceInput input;

    private void Awake()
    {
        var coherenceSync = GetComponent<CoherenceSync>();
        fixedUpdateInput = coherenceSync.MonoBridge.FixedUpdateInput;
        input = coherenceSync.Input;
    }

    // Retrieves the "jump" input state for a given frame
    public bool GetJump(long frame)
    {
        return input.GetButtonState("Jump", frame);
    }

    // Sets the "jump" state for the current frame
    public void SetJump()
    {
        bool hasJumped = fixedUpdateInput.GetKey(KeyCode.Space);
        input.SetButtonState("Jump", hasJumped);
    }
}

The FixedUpdateInput works by sampling inputs at Update and prolonging their lifetime to the network FixedNetworkUpdate so they can be processed correctly there. For our last example that would mean "down" & "pressed" registered in the first FixedNetworkUpdate after the initial five updates, followed by an "up" state in the subsequent FixedNetworkUpdate.

Pause handling

There's a limit to how many frames can be predicted by the clients. This limit is controlled by the CoherenceInput.InputBufferSize. When clients try to predict too many frames into the future (more frames than the size of the buffer) the simulation will issue a pause. This pause affects only the local client. As soon as the client receives enough inputs to run another frame the simulation will resume.

To get notified about the pause use the OnPauseChange(bool isPaused) method from the CoherenceInputSimulation:

public class Simulation : CoherenceInputSimulation<SimulationState>
{
    // ... Simulation implementation ...
    
    protected override void OnPauseChange(bool isPaused)
    {
        PauseScreen.SetActive(isPaused);
    }
}

This can be used for example to display a pause screen the informs user about a bad internet connection.

Debugging

The CoherenceInputSimulation has a built-in debugging utility that collects various information about the input simulation on each frame. This data can prove extremely helpful in finding a simulation desync point.

Setup

Since debugging might induce a non-negligible overhead it is turned off by default. To turn it on add, a COHERENCE_INPUT_DEBUG scripting define:

From that point, all the debugging information will be gathered. The debug data is dumped to a JSON file as soon as the client disconnects. The file can be located under a root directory of the executable (in case of Unity Editor the project root directory) under the following name: inputDbg_<ClientId>.json, where <ClientId> is the CoherenceClientConnection.ClientId of the local client.

Data handling behaviour can be overridden by setting the CoherenceInputDebugger.OnDump delegate, where the string parameter is a JSON dump of the data.

The debugger is available as a property in the simulation base class: CoherenceInputSimulation.Debugger. Most of the debugging data is recorded automatically, however, the user is free to append any arbitrary information to a frame debug data, as long as it is JSON serializable. This is done by using the CoherenceInputDebugger.AddEvent method:

public class Simulation : CoherenceInputSimulation<SimulationState>
{
    // ... Simulation implementation ...
    
    protected override void SetInputs(CoherenceClientConnection client)
    {
        Debugger.AddEvent("myCustomEvent", new
        {
            Data = "my custom data",
            UnityFrameCount = Time.frameCount
        });
        // ... movement code ...
    }
}

Since the simulation can span an indefinite amount of frames it might be wise to limit the number of debug frames kept by the debugging tool (it's unlimited by default). To do this use the CoherenceInputDebugger.FramesToKeep property. For example, setting it to 1000 will instruct the debugger to keep only the latest 1000 frames worth of debugging information in the memory.

Serialization

Since the debugging tool uses JSON as a serialization format, any data that is part of the debug dump must be JSON-serializable. An example of this is the simulation state. The simulation state from the quickstart example is not JSON serializable by default, due to Unity's Vector3 that doesn't serialize well out of the box. To fix this we need to give JSON serializer a hint:

public struct SimulationState : IHashable
{
    [JsonProperty(ItemConverterType = typeof(UnityVector3Converter))]
    public Vector3[] PlayerPositions { get; set; }
}

With the JsonProperty attribute, we can control how a given field/property/class will be serialized. In this case, we've instructed the JSON serializer to use the custom UnityVector3Converter for serializing the vectors.

Comparing debug data

To find a problem in the simulation, we can compare the debug dumps from multiple clients. The easiest way to find a divergence point is to search for a frame where the hash differs for one or more of the clients. From there one can inspect the inputs and simulation states from previous frames to find the source of the problem.

Here's the debug data dump example for one frame:

"32625635555": {
    "Frame": 32625635555,
    "AckFrame": 32625635556,
    "ReceiveFrame": 32625635556,
    "AckedAt": 32625635555,
    "MispredictionFrame": -1,
    "Hash": "3fa00ad32cee971e43ee4a3206dc79f0",
    "Time": "15:48:36.977",
    "InitialState": {
      "PlayerPositions": [
        "0.2494998,3.992,0",
        "0.2494998,-2.086163E-07,0",
        "-2.086163E-07,0.7484999,0",
        "0.4989998,3.992,0"
      ]
    },
    "InitialInputs": {
      "153470363": "Mov:float2(-0.998f, 0f)",
      "322232370": "Mov: float2(0.998f, 0f)",
    },
    "InputBufferStates": {
      "153470363": {
        "LastFrame": 32625635557,
        "LastSentFrame": 32625635557,
        "LastReceivedFrame": -1,
        "LastAcknowledgedFrame": -1,
        "MispredictionFrame": null,
        "QueueCount": 0,
        "ShouldPause": false
      },
      "322232370": {
        "LastFrame": 32625635556,
        "LastSentFrame": -1,
        "LastReceivedFrame": 32625635556,
        "LastAcknowledgedFrame": 32625635556,
        "MispredictionFrame": null,
        "QueueCount": 0,
        "ShouldPause": false
      }
    },
    "Events": [
      {
        "Event": "InputSent",
        "Time": "15:48:36.977",
        "Data": {
          "ClientId": 153470363,
          "SentFrame": 32625635558,
          "Input": "Mov:float2(0f, -0.998f)"
        }
      },
      {
        "Event": "InputReceived",
        "Time": "15:48:36.978",
        "Data": {
          "ClientId": 322232370,
          "RecvFrame": 32625635557,
          "Input": "Mov: float2(0.998f, 0f)"
        }
      }
    ]
  },

Explanation of the fields:

• Frame - frame of this debug data

• AckFrame - the common acknowledged frame, i.e. the lowest frame for which inputs from all clients have been received and are known to be valid (not mispredicted)

• ReceiveFrame - the common received frame, i.e. the lowest frame for which inputs from all clients have been received

• AckedAt - a frame at which this frame has been acknowledged, i.e. set as known to be valid (not mispredicted)

• MispredictionFrame - a frame that is known to be mispredicted, or -1 if there's no misprediction

• Hash - hash of the simulation state. Available only if the simulation state implements the IHashable interface

• Initial state - the original simulation state at this frame, i.e. a one before rollback and resimulation

• Initial inputs - original inputs at this frame, i.e. ones that were used for the first simulation of this frame

• Updated state - the state of the simulation after rollback and resimulation. Available only in case of rollback and resimulation

• Updated inputs - inputs after being corrected (post misprediction). Available only in case of rollback and resimulation

• Input buffer states - dump of the input buffer states for each client. For details on the fields see the InputBuffer code documentation

• Events - all debug events registered in this frame

State hash

One of the things that are stored as part of the debugging information is the simulation state. In the case of a complex state searching for a difference across multiple clients and hundreds of frames can quickly become tedious. To simplify the problem we can use the hash calculation feature of the input debugger.

Every state simulation class/struct that implements the IHashable interface will have its hash automatically calculated and stored as part of the debugging information. The example of IHashable implementation:

public struct SimulationState : IHashable
{
    [JsonProperty(ItemConverterType = typeof(UnityVector3Converter))]
    public Vector3[] PlayerPositions { get; set; }

    public Hash128 ComputeHash()
    {
        return Hash128.Compute(PlayerPositions);
    }
}

Configuration

Input buffer

There are two main variables which affect the behaviour of the InputBuffer:

• Input buffer size - the size of the buffer determines how far into the future the input system is allowed to predict. The bigger the size, the more frames can be predicted without running into a pause. Note that the further we predict, the more unexpected the rollback can be for the player. The InitialBufferSize value can be set directly in code however it must be done before the Awake of the baked component, which might require a script execution order configuration.

• Input buffer delay - dictates how many frames must pass before applying an input. In other words, it defines how "laggy" the input is. The higher the value, the less likely clients are going to run into prediction (because a "future" input is sent to other clients), but the more unresponsive the game might feel. This value can be changed freely at runtime, even during a simulation (it is however not recommended due to inconsistent input feeling).

The other two options are:

• Disconnect on time reset - if set to 'true' the input system will automatically issue a disconnect on an attempt to resync time with the server. This happens when the client's connection was so unstable that frame-wise it drifted too far away from the server. In order to recover from that situation, the client performs an immediate "jump" to what it thinks is the actual server frame. There's no easy way to recover from such "jump" in the deterministic simulation code so the advised action is to simply disconnect.

• Use fixed simulation frames - if set to 'true' the input system will use the IClient.ClientFixedSimulationFrame frame for simulation - otherwise the IClient.ClientSimulationFrame is used. Setting this to 'true' is recommended for the deterministic simulation.

Fixed frame rate

The fixed network update rate is based on the Fixed Timestep configured through the Unity project settings:

To know the exact fixed frame number that is executing at any given moment use the IClient.ClientFixedSimulationFrame or CoherenceInputSimulation.CurrentSimulationFrame property.

Overview

Commands are network messages sent from one entity to another entity in the networked world. Functionally equivalent to standard RPCs, commands bind to public methods accessible on the GameObject that CoherenceSync sits on.

Design Phase

In the design phase, you can expose public methods the same way you select fields for synchronization: through the Configure window on your CoherenceSync component.

Selected public methods will be exposed as network commands in the baking process.

The button on the right of the method lets you choose the routing mode. Commands with aSend to Authority Only mode can be sent only to the authority of the target entity, while ones with the Send to All Instances can be broadcasted to all clients that have a copy of this entity. The routing is enforced by the Replication Server as a security measure so the outdated or malicious clients don't break the game.

Sending a Command on a Networked Object

To send a command, we call the SendCommand __ method on the target CoherenceSync object. It takes a number of arguments:

• The type argument (within the <and>) must be the type of the receiving MonoBehaviour. This ensures that the correct method gets called if the receiving GameObject has components that implement methods that share the same name.

• The first argument is the name of the method on the MonoBehaviour that we want to call. It is good practice to use the C# nameof expression when referring to the method name, since it prevents accidentally misspelling it, or forgetting to update the string if the method changes name.

• The second argument is an enum that specifies the MessageTarget of the command. The possible values are:

  • MessageTarget.All – this will send the command to each client that has an instance of this entity.

  • MessageTarget.AuthorityOnly – this will send the command only to the client that has authority over the entity.

  Mind that the target must be compatible with the routing mode set in the bindings, i.e. Send to authority will allow only for the MessageTarget.AuthorityOnly while Send to all instances allows for both values.

• The rest of the arguments (if any) vary depending on the command itself. We must supply as many parameters as are defined in the target method and the schema.

Here's an example of how to send a command:

var target = enemy.GetComponent<CoherenceSync>();
                    
target.SendCommand<Character>(nameof(SendDamage), 
                MessageTarget.All,
                damageValue, staminaBlockCost,
                staminaRecoveryDelay, ignoreDefense, 
                hitReaction, hitPosition, 
                ConnectionId, isPlayer);

Receiving a Command

We don't have to do anything special to receive the command. The system will simply call the corresponding method on the target network entity.

If the target is a locally simulated entity, SendCommand will recognize that and not send a network command, but instead simply call the method directly.

Sending a Command to multiple entities

Sometimes you want to inform a bunch of different entities about a change. For example, an explosion impact on a few players. To do so, we have to go through the instances we want to notify, and send commands to each of them.

using Coherence;
using Coherence.Toolkit;
using UnityEngine;

public class MyCommands : MonoBehaviour
{
    public void MulticastCommand()
    {
        var self = GetComponent<CoherenceSync>();
        var listeners = FindObjectsOfType<CoherenceSync>(); 
        
        foreach (var listener in listeners)
        {
            if (!listener || listener == self)
            {
                continue;
            }
            
            listener.SendCommand<MyCommands>(nameof(ReceiveCommand), MessageTarget.AuthorityOnly);
        }
    }

// bind to this method via the Bindings window
    public void ReceiveCommand()
    {
        Debug.Log("Received!");
    }
}

In this example, a command will get sent to each network entity under the authority of this client. To make it only affect entities within certain criteria, you need to filter to which CoherenceSync you send the command to on your own.

Sending null values in command arguments

Some of the primitive types supported are nullable values, this includes:

• Byte[]

• string

• Entity references: CoherenceSync, Transform, and GameObject

In order to send one of these values as a null (or default) we need to use special syntax to ensure the right method signature is resolved.

using Coherence;
using Coherence.Toolkit;
using UnityEngine;

public class MyCommand : MonoBehaviour
{
    public void MyNullableCommand(string someString, Byte[] someArray)
    {
        if (!string.IsNullOrEmpty(someString)) //Could be null or string.Empty
        {
            // Do something with the string
        }
        
        if (someArray != null && someArray.Length > 0) //Could be null or Byte[0]
        {
            // Do something with the array
        }
    }
}

sync.SendCommand<MyCommand>(nameof(MyNullableCommand), 
    MessageTarget.All, 
    (typeof(string), (string)null), 
    (typeof(Byte[]), (Byte[])null));

Null-value arguments need to be passed as a ValueTuple<Type, object> so that their type can be correctly resolved. In the example above sending a null value for a string is written as: (typeof(string), (string)null)

and the null Byte[] argument is written as: (typeof(Byte[]), (Byte[])null)

Mis-ordered arguments, type mis-match, or unresolvable types will result in errors logged and the command not being sent.

Supported types in Commands

When a prefab is not using a baked script there are some restrictions for what types can be sent in a single command:

• 4 entity references

• maximum of 511 bytes total of data in other arguments

• a single Byte[] argument can be no longer than 509 bytes because of overhead

Some network primitive types send extra data when serialized (like Byte arrays and string types) and some are compressed using default compression settings (like int, float, Vector2, and Vector3) so gauging how many bits a command will use is difficult. If a single command is bigger than the supported packet size, it won't work even with baked code. For a good and performant game experience, always try to keep the total command argument sizes low.

Unity Animator's parameters are bindable out of the box, with the exception of triggers.

Triggers

Triggers can be invoked over the network using commands. Here's an example where we inform networked clients that we have played a jump animation:

using UnityEngine;
using Coherence;
using Coherence.Toolkit;
using System.Collections.Generic;

public class JumpController : MonoBehaviour
{
    CoherenceSync coherenceSync;
    Animator animator;

    void Awake()
    {
        coherenceSync = GetComponent<CoherenceSync>();
        animator = GetComponent<Animator>();
    }

    void Update()
    {
        if (!coherenceSync.isSimulated)
        {
            return;
        }

        if (Input.GetKeyDown(KeyCode.Space))
        {
            MakePlayerJump();
        }
    }

    void MakePlayerJump()
    {
        coherenceSync.SendCommand<JumpController>(nameof(PlayJumpAnimation), MessageTarget.All, coherenceSync);
    }

    // bind to this method via the Bindings window
    public void PlayJumpAnimation(CoherenceSync jumpSync)
    {
        animator.SetTrigger("Jump");
    }
}

Now, bind to PlayJumpAnimator.

Overview

The client connection system lets you uniquely identify users connected to the same session, find any user by his ID, spawn objects whenever a new user joins the session, and send messages between those users.

To achieve this a special connection entity is automatically created by the replication server for each connected client, including a simulator. Those entities are subject to a different rule set than standard entities. Connection entities:

• Can't be created or destroyed by the client - they are always replication server-driven

• Are global - they are replicated across clients regardless of the in-simulation distance or LiveQuery radius

Client connections shine whenever there's a need to communicate something to all the connected players. Usage examples:

• Global chat

• Game state changes: game started, game ended, map changed

• Server announcements

• Server-wide leaderboard

• Server-wide events

Enabling client connections

The globality of client connection doesn't fit all game types - for example, it rarely makes sense to keep every client informed about the presence of all players on the server in an MMORPG game (think World Of Warcraft). Due to this client connections are turned off by default.

To enable client connections turn on the global query in the MonoBridge (the Global Query On toggle):

Connection management

Most of the client connection functionality is accessible through the CoherenceMonoBridge.ClientConnections object:

using System.Collections.Generic;
using Coherence.Connection;
using Coherence.Toolkit;
using UnityEngine;

public class GameManager : NetBehaviour
{
    public CoherenceMonoBridge MonoBridge;

    void Start()
    {   
        // IMPORTANT: All of the connection retrieving calls may return null 
        // if the connection system was not turned on, not initialized yet,
        // or simply if the connection was not found.

        // Raised whenever a new connection is made (including the local one).
        MonoBridge.ClientConnections.OnCreated += connection =>
        {
            Debug.Log($"Connection #{connection.ClientId} " +
                      $"of type {connection.Type} created.");
        };
        
        // Raised whenever a connection is destroyed.
        MonoBridge.ClientConnections.OnDestroyed += connection =>
        {
            Debug.Log($"Connection #{connection.ClientId} " +
                      $"of type {connection.Type} destroyed.");
        };

        // Specifies how many clients are in this session (room or world).
        int clientCount = MonoBridge.ClientConnections.ClientConnectionCount;

        // Returns connection objects for all connections in this session.
        IEnumerable<CoherenceClientConnection> allConnections 
            = MonoBridge.ClientConnections.GetAll();

        // Returns all connections except for the local one.
        IEnumerable<CoherenceClientConnection> otherConnections 
            = MonoBridge.ClientConnections.GetOther();

        // Returns connection object of the local user.
        CoherenceClientConnection myConnection 
            = MonoBridge.ClientConnections.GetMine();
        
        // Returns connection object of the simulator (if one is connected).
        CoherenceClientConnection simulatorConnection 
            = MonoBridge.ClientConnections.GetSimulator();

        // Retrieves a connection by its ClientID.
        CoherenceClientConnection selectedConnection 
            = MonoBridge.ClientConnections.Get(myConnection.ClientId);
        
        // Retrieves a connection by its EntityID (warning: requires
        // connection prefab with a CoherenceSync attached).
        if (myConnection.Sync != null)
        {
            selectedConnection 
                = MonoBridge.ClientConnections.Get(myConnection.Sync.EntityID);
        }
        
        // Specifies if this is a client or a simulator connection.
        ConnectionType connectionType = selectedConnection.Type;
        
        // Specifies if this is a local connection (belonging to the
        // local user).
        bool isMine = selectedConnection.IsMyConnection;
        
        // Returns a GameObject associated with this connection.
        // Applicable only if connection prefabs are used.
        GameObject connectionGameObject = selectedConnection.GameObject;
    }
}

Each connection is represented by a plain C# CoherenceClientConnection object. It contains all the important information about a connection - its ClientID, Type, whether it IsMyConnection, and a reference to the GameObject and Coherence Sync associated with it.

Connection objects

Each client connection can have a GameObject with CoherenceSync automatically being spawned and associated with it. Those objects, like any other objects with CoherenceSync, can be used for syncing properties or sending messages, with a little twist - they are global and thus not limited by the live query radius. That makes them perfect candidates for operations like:

• Syncing global information - name, stats, tags, etc.

• Sending global messages - chat, server interaction

To enable connection objects:

1. Create a client connection prefab

This step is described in detail in the Prefab setup section. In short, a prefab with a CoherenceSync and a custom component (PlayerConnection in this example) must be created and placed in the Resources directory:

2. Link a connection prefab to the MonoBridge

For the system to know which object to create for every new client connection, we have to link our prefab to the MonoBridge. Simply drag the prefab to the Client field in the MonoBridge inspector:

From now on every new connection will be assigned an instance of this prefab, which can be accessed through the CoherenceClientConnection.GameObject property.

Prefab selection

Note that there's a separate field for the Simulator Connection Prefab. It can be used to spawn a completely different object for the simulator connection that may contain simulator specific commands and replicated properties. If the field is left empty, no object will be created for the simulator connection.

The prefab selection process can be also controlled from code using the CoherenceMonoBridge.ClientConnections.ProvidePrefab callback:

using System.Collections.Generic;
using Coherence.Connection;
using Coherence.Toolkit;
using UnityEngine;

public class GameManager : NetBehaviour
{
    public CoherenceMonoBridge MonoBridge;
    public CoherenceSync CustomConnectionPrefab;

    void Start()
    {
        MonoBridge = FindObjectOfType<CoherenceMonoBridge>();
        MonoBridge.ClientConnections.ProvidePrefab += (clientId, connectionType) =>
        {
            return CustomConnectionPrefab;
        };
    }
}

Client messages

Client messages are commands sent between the client connection objects. Implementing client messages is as simple as adding a new method to the component used by our connection prefab and binding it in the configuration:

public class PlayerConnection : MonoBehaviour
{
    // My chat command
    public void OnChatMessage(string message)
    {
        Debug.Log($"Received chat message: {message}");
    }
}

Don't forget to bind the new command:

Client messages can be sent using the CoherenceClientConnection.SendClientMessage method:

using Coherence;
using Coherence.Toolkit;
using UnityEngine;

public class PlayerConnection : MonoBehaviour
{
    void Start()
    {
        var monoBridge = FindObjectOfType<CoherenceMonoBridge>();
        CoherenceClientConnection myConnection = monoBridge.ClientConnections.GetMine();
        
        myConnection.SendClientMessage<PlayerConnection>(nameof(OnChatMessage), 
            MessageTarget.Other, "Hello!");
    }
    
    // My chat command
    public void OnChatMessage(string message)
    {
        Debug.Log($"Received chat message: {message}");
    }
}

If the ClientID of the message recipient is known we can use the CoherenceMonoBridge.ClientConnections directly to send a client message:

private int lastClientId;

private void OnConnectionCreated(CoherenceClientConnection clientConnection)
{
    if (!clientConnection.IsMyConnection)
    {
        lastClientId = clientConnection.ClientId;
    }
}

public void SendMessageToLastConnection(string message)
{
    MonoBridge.ClientConnections.SendMessage<PlayerConnection>("OnChatMessage",
        lastClientId, MessageTarget.AuthorityOnly, "Hello!");
}

What are Input Queues?

Input queues enable a simulator to take control of the simulation of another client's objects based on the client's inputs.

When To Use Input Queues?

• In situations where you want a centralized simulation of all inputs. Many game genres use input queues and centralized simulation to guarantee the fairness of actions or the stability of physics simulations.

• In situations where clients have low processing power. If the clients don't have sufficient processing power to simulate the world it makes sense to use input queue and just display the replicated results on the clients.

• In situations where determinism is important. RTS and fighting games will use input queues and rollback to process input events in a shared (not centralized), and deterministic, way so that all clients simulate the same conditions and produce the same results.

Setup With CoherenceSync and CoherenceInput

Setting up an object to simulate via input queues using CoherenceSync is done in three steps:

1. Preparing the CoherenceSync component on the object prefab

The Simulation Type of the CoherenceSync component is set to Simulation Server With Client Input

Setting the simulation type to this mode instructs the client to automatically give authority over this object to the simulator in charge of simulating all inputs on all objects.

2. Declaring Inputs for the simulated object

Each simulated CoherenceSync component is able to define its own, unique set of inputs for simulating that object. An input can be one of:

• Button. A button input is tracked with just a binary on/off state.

• Button Range. A button range input is tracked with a float value from 0 to 1.

• Axis. An axis input is tracked as two floats from -1 to 1 in both the X and Y axis.

• String. A string value representing custom input state. (max length of 63 characters)

To declare the inputs used by the CoherenceSync component, the CoherenceInput component is added to the object. The input is named and the fields are defined.

In this example, the input block is named "Player Movement" and the inputs are WASD and "mouse" for the XY mouse position.

3. Bake the CoherenceSync object

In order for the inputs to be simulated on CoherenceSync objects, they must be optimized through baking.

Using CoherenceInput

When a simulator is running it will find objects that are set up using CoherenceInput components and will automatically assume authority and perform simulations. Both the client and simulator need to access the inputs of the CoherenceInput of the replicated object. The client uses the Set* methods and the simulator uses the Get* methods to access the state of the inputs of the object. In all of these methods, the name parameter is the same as the Name field in the CoherenceInput component.

Client-Side Set* Methods

• public void SetButtonState(string name, bool value)

• public void SetButtonRangeState(string name, float value)

• public void SetAxisState(string name, Vector2 value)

• public void SetStringState(string name, string value)

Simulator-Side Get* Methods

• public bool GetButtonState(string name)

• public float GetButtonRangeState(string name)

• public Vector2 GetAxisState(string name)

• public string GetStringState(string name)

For example, the mouse click position can be passed from the client to the simulator via the "mouse" field in the setup example.

using Coherence.Toolkit;
using UnityEngine;

public void SendMousePosition()
{
    var coherenceInput = GetComponent<CoherenceInput>();
    var mousePos = Input.mousePosition;
    coherenceInput.SetAxisState("mouse", new Vector2(mousePos.x, mousePos.y));
}

The simulator can access the state of the input to perform simulations on the object which are then reflected back to the client as any replicated object is.

using Coherence.Toolkit;
using UnityEngine;

public void ProcessMousePosition()
{
    var coherenceInput = GetComponent<CoherenceInput>();
    var mousePos = coherenceInput.GetAxisState("mouse");
    
    //Move object
}

Input Owner

Each object only accepts inputs from one specific client, called the object's Input Owner.

When a client spawns an object it automatically becomes the Input Owner for that object. This way, the object's creator will retain control over the object even after authority has been transferred to the simulator.

If an object is spawned directly by the simulator, you will need to assign the Input Owner manually. Use the SetInputOwner method on the CoherenceInput component to assign or re-assign a client that will take control of the object:

public void AssignNewInputOwner(ClientConnection newInputOwner)
{
    var coherenceInput = GetComponent<CoherenceInput>();
    coherenceInput.SetInputOwner(newInputOwner.ClientId);
}

The ClientId used to specify input owner can currently only be accessed from the ClientConnection class. For detailed information about setting up the ClientConnection prefab, see the Client connections page.

Client-Side Prediction

Generally, using server-side simulation with simulators makes the time from the client providing input to the time the object is updated with that input significantly longer than just client-side simulation because of the time required for the input to be sent to the simulator, processed, and then the updates to the object returned across the network. This can cause visual lag. Using input queues allows the client to perform prediction of the simulation to provide a smooth playing experience.

If the client simulates the inputs on the object as well as applying them to the CoherenceInput component and then blends the authoritative results from the simulator with the locally simulated results, then smoother and more responsive simulation is achieved.

Lifetime configuration

The CoherenceSync editor interface allows us to define the Lifetime of a networked object. The following options are available:

• Session Based. No persistence. The entity will disappear when the client or simulator disconnects.

• Persistent. The entity will remain on the server until a simulating client deletes it.

Uniqueness and UUID

Unique persistent objects need to be identified so that the system can know how to treat duplicate persistent objects.

Manually assigning a UUID means that each instance of this persistent object prefab is considered the same object regardless of where on the network it is instantiated. So, for example, if two clients instantiate the same prefab object with the same persistence UUID then only one is considered official and the other is replaced by the replication server.

The CoherenceUUID behaviour is used to uniquely identify a prefab.

It has several functions: you can generate a new ID for your object, and you can set auto-generate UUID on the scene to true, so each time object will receive a new ID.

Deleting a persistent object

Deleting a persistent object is done the same as with any network object - by destroying its GameObject.

Bandwidth is limited

No matter how fast the internet becomes, conserving bandwidth will always be important. Some game clients might be on poor mobile networks with low upload and download speeds, or have high ping to the replication server and/or other clients, etc.

Additionally, sending more data than is required consumes more memory and unnecessarily burdens the CPU and potentially GPU, which could add to performance issues, and even to quicker battery drainage.

Optimization techniques

In order to optimize the data we are sending over the network, we can employ various techniques built into the core of coherence.

• Delta-compression (automatic). When possible, only send differences in data, not the entire state every frame.

• Compression and quantization (automatic and configurable). Various data types can be compressed to consume less bandwidth that they naturally would.

• Simulation frequency (configurable). Most entities do not need to be simulated at 60+ frames per second.

• Levels of detail (configurable). Entities need to consume less and less bandwidth the farther away they move from the observer.

• Area of interest. Only replicate what we can see.

LiveQuery

The way you get information about the world is through LiveQueries. We set criteria for what part of the world we are interested in at each given moment. That way, the replicator won’t send information about everything that is going on in the game world everywhere, at all times.

Instead, we will just get information about what’s within a certain area, kind of like moving a torch around to look in a dark cave.

Adding a LiveQuery to the scene

A LiveQuery is a cube that defined the area of interested in a particular part of the world. It is defined by its position and its radius (half the side of the cube). There can be multiple LiveQueries in a single scene.

Moving a LiveQuery

A classic approach is to put a LiveQuery on the camera and set the radius to correspond to the far clipping plane or visibility distance.

Moving the GameObject containing the LiveQuery will also notify the replication server that the query for that particular game client has moved.

Adding a TagQuery

In addition to the LiveQuery, coherence also supports filtering objects by tag. This is useful when you have some special objects that should always be visible regardless of world position.

To create a TagQuery, right click a GameObject in the scene and select coherence -> TagQuery from the context menu.

All networked GameObjects with matching tags will now be visible to the client. The coherence tag can be any string and can be configured separately from the Unity tag in the Advanced Settings section of the CoherenceSync component.

Tags and TagQueries can be updated at any time while the application is running, either from the Unity inspector or setting CoherenceSync.tag and CoherenceTagQuery.tag with code.

Currently, only a single tag per GameObject and TagQuery is supported. To include objects with different tags, you can create multiple TagQuery objects for each tag.

If we don't do any special configuration, entity data is captured at the highest possible frequency and sent to the replication server. This often generates more data than is needed to efficiently replicate the entity's state across the network.

Global simulator frequency

On a simulator, we can limit the framerate globally using Unity's built-in static variable targetFrameRate.

Per-component frequency limitation

Sample rate can also be configured individually for all fields with code.

This document explains how to set up an ever increasing counter that all clients have access to. This could be used to make sure that everyone can generate unique identifiers, with no chance of ever getting a duplicate.

By being persistent, the counter will also keep its value even if all clients log off, as long as the replication server is running.

The Counter

First, create a script called Counter.cs and add the following code to it:

Next, add this script to a prefab with CoherenceSync on it, and select the counterand the method NextNumber for syncing in the bindings window. To make the counter behave like we want, mark the prefab as "Persistent" and give it a unique persistence ID, e.g. "THE_COUNTER". Also change the adoption behaviour to "Auto Adopt":

Finally, make sure that a single instance of this prefab is placed in the scene.

NumberRequester

Now, create a script called NumberRequester.cs. This will be an example MonoBehaviour that requests a unique number by sending the command GetNumber to the Counter prefab. As a single argument to this command, the NumberRequester will send an entity reference to itself. This makes it possible for the Counter to send back a response command (GotNumber) with the number that was generated. In this simple example we just log the number to the console.

To make this script work, add it to a prefab that has the CoherenceSync script and mark the GotNumber for syncing in the bindings window.

Sometimes you want to synchronize entities that are connected to other entities. These relationships can be references between entities, but they can also involve direct parent-child relationship between game objects, or more nuanced use cases.

Here's a guide for what technique to use, depending on the situation:

• If you have an entity that needs to keep a nullable reference to another entity, use a normal Entity reference. This includes any existing MonoBehaviour that has GameObject or Transform fields that you want to synchronize over the network.

• If the entities are placed in a hierarchy, use the techniques for parent-child relationships.

Extrapolation or "dead reckoning" uses historical data to predict the future state of a component. The actual network data that arrives later can be used to interpolate or snap the predicted values to the correct ones.

Why does world size matter?

Positions and livequeries in the world are compressed, and the compression is defined by maximum scale and bit count. Larger scale at the same bit count means lower precision, and vice-versa.

Defining world size

A default maximum world size of 2400 means that only values from [-2400, 2400] will be supported for all spatial axes.

If our game scenes are larger than 2400 x 2 (4800) units across, we can increase this value in the Settings window or in the schema as illustrated below.

Don't forget to extend the LiveQuery scale to match the world position scale.

# Default world positions:

component WorldPosition
  value Vector3 [bits "24", scale "5000"]
  

# LiveQuery:

component WorldPositionQuery
  position Vector3 [bits "24", scale "5000"]
  radius Float [bits "24", scale "5000"]

Overview

Depending on the settings in our project, data may not always arrive at a smooth 60 frames per second through the network. This is completely okay, but in order to make state changes (e.g. movement, rotation) appear smooth on the client, we use interpolation.

Interpolation is a type of estimation, a method of constructing new data points within the range of a discrete set of known data points.

The way interpolation works in coherence is that we wait for three data points and then start smoothing the subsequent values according to the interpolation parameters defined in the interpolation settings provided.

Binding Interpolation Settings

In the Configure window, each binding displays its interpolation settings next to it.

Built-in interpolation settings for position and rotation are provided out-of-the-box, but you are free to create your own and use them instead.

There, you have a few settings you can tweak:

• Interpolation Type: the type of interpolation used. If set to None, the value will simply snap to the closest sample point without any blending.

• Smooth Time: additional smoothing can be applied (using SmoothDamp) to clear out any jerky movement after regular interpolation has been performed.

• Max Smoothing Speed: the maximum speed at which the value can change, unless teleporting.

• Teleport Distance: if two consecutive samples are further apart than this, the value will "teleport" or snap to the new sample immediately without interpolating or smoothing in between.

• Latency Control: Auto

  • With Latency Control set to Auto, latency will automatically adapt to the average sample delta time for all samples in the buffer. This allows each binding to operate with minimal latency without having to extrapolate data.

  • Factor: fudge factor applied to the average sample delta time. A factor of 1 means latency is exactly equal to time between samples, so the interpolated value should reach the last sample in the buffer at the exact time when a new sample is expected to arrive.

    • In general, a factor of 1.1 is recommended to keep from running out of samples (due to network fluctuations).

    • For spline interpolation, a factor of at least 2 is recommended, because the spline algorithm requires staying two samples behind to produce smooth curves.

• Latency Control: Manual

  • With Latency Control set to Manual, latency is fixed so the interpolated value always stays a certain time behind the game time.

  • Latency: seconds the interpolated value will trail being the game time.

• Overshooting

  • Max: how far into dead reckoning to venture when the time fraction exceeds 100%, as percentage of the sample rate.

  • Retraction: how fast to pull back to 100% when overshooting the allowed dead reckoning maximum (in seconds)

Interpolation works both in baked and reflection modes. You can change these settings at runtime via the Configure window (editor) or accessing the binding and changing the interpolation settings yourself:

if (coherenceSync.TryGetBinding(typeof(Transform), "position", out IBinding binding)
{
    // change your interpolation settings at runtime
    binding.InterpolationSettings = ...
}

Entity references let you set up references between entities and have those be synchronized, just like other value types (like integers, vectors, etc.)

To use Entity references, simply select any fields of type GameObject, Transform, or CoherenceSync for syncing in the Configuration window:

The synchronization works both when using reflection and in baked sync scripts.

Entity references can also be used as arguments in Commands.

Limitations

It's important to know about the situations when an entity reference might become null, even though it seems like it should have a value:

• A client might not have the referenced entity in its live query. A local reference can only be valid if there's an actual entity instance to reference. If this becomes a problem, consider switching to using the ConnectedEntity component which ensures that another entity becomes part of the query.

• The owner of the entity reference might sync the reference to the Replication Server before syncing the referenced entity. This will lead to the Replication Server storing a null reference. If possible, try setting the entity references during gameplay when the referenced entities have already existed for a while.

In any case, it's important to use a defensive coding style when working with entity references. Make sure that your code can handle missing entities and nulls in a graceful way.

Motivation

coherence can support large game worlds with many objects. Since the amount of data that can be transmitted over the network is limited, it's very important to only send the most important things.

You already know a very efficient tool for enabling this – the LiveQuery. It ensures that a client is only sent data when an object in its vicinity has been updated.

Often though, there is a possibility for an even more nuanced and optimized approach. It is based on the fact that we might not need to send as much data for an entity that is far away, compared to a close one. A similar technique is often used in 3D-programming to show a simpler model when something is far away, and a more detailed when close-up.

This idea works really well for networking too. For example, when another player is close to you it's important to know exactly what animation it is playing, what it's carrying around, etc. When the same player is far off in the horizon, it might suffice to only know it's position and orientation, since nothing else will be discernible anyways.

To use this technique we must learn about something called archetypes.

Archetypes

An archetype is a component that can be added to any prefab with the CoherenceSync component. It contains a list of the various levels of detail (LODs) that this particular prefab can have.

There must always exist a LOD 0, this is the default level and it always has all components enabled (it can have per-field overrides though, see below.)

There can be any number of subsequent LODs (e.g. LOD 1, LOD 2, etc.) and each one must have a distance threshold higher than the previous one. The coherence SDK will try to use the LOD with the highest number, but that is still within the distance threshold.

One each LOD, there are two options for optimizing data being transferred:

1. Components can be turned off, meaning you won't receive any updates from it.

2. It's fields can be configured to use fewer bits, usually leading to less fine-grained information. The idea is that this won't be noticeable at the distance of the LOD.