Introduction

There are a large number of communication protocols used by OpenSimulator. These can be divided into the four types.

Client protocols

These are communication protocols between the server and a client/viewer. The primary protocol here is the Linden Lab viewer protocol. This protocol is divided into UDP messages (such as object updates, avatar position updates, etc.) and HTTP based messages via capabilities.

The Second Life wiki has an incomplete summary of current capabilities.

There are also separate pages for

• Inventory capabilities (FetchInventoryDescendents2, FetchInventory2, FetchLibDescendents2, FetchLib2).
• Texture capabilities (GetTexture).
• SimulatorLoginProtocol

Grid service protocols

On a standalone OpenSimulator installation, all communications occurs within process. However, with a grid installation the simulators need to communicate with backend services (asset, inventory, etc.). This is done over HTTP.

Simulator to simulator protocols

There are some situations in which simulators need to communicate directly with one another. One of these is on teleport or region crossing. This communication is carried out through HTTP.

This section should also include archive formats (e.g. OARs and IARs). This encompasses a very large number of object and asset formats (e.g. animations, textures).

Simulator to external protocols

There are some ways in which the simulator can be examined or controlled externally. One example of control is the RemoteAdmin interface. There are also facilities for Monitoring the current state of the simulator.

References

There's a list of various OpenSimulator entry points at the page Known Web Interfaces within OpenSim but this is probably pretty out of date.