configuration.md

General

This file describes some of the java properties which ice4j uses to configure itself.

Interfaces and IP addresses

org.ice4j.ice.harvest.ALLOWED_INTERFACES

Default: all interfaces are allowed.

This property can be used to specify a ";"-separated list of interfaces which are allowed to be used for candidate allocations. If not specified, all interfaces are considered allowed, unless they are explicitly blocked (see below).

org.ice4j.ice.harvest.BLOCKED_INTERFACES

Default: no interfaces are blocked.

This property can be used to specify a ";"-separated list of interfaces which are not allowed to be used for candidate allocations.

org.ice4j.ice.harvest.ALLOWED_ADDRESSES

Default: all addresses are allowed.

This property can be used to specify a ";"-separated list of IP addresses which are allowed to be used for candidate allocations. If not specified, all addresses are considered allowed, unless they are explicitly blocked (see below).

org.ice4j.ice.harvest.BLOCKED_ADDRESSES

Default: no addresses are blocked.

This property can be used to specify a ";"-separated list of IP addresses which are not allowed to be used for candidate allocations.

org.ice4j.ipv6.DISABLED

Type: boolean

Default: false

This property can be used to disable binding on IPv6 addresses.

Mapping harvesters

Ice4j uses the concept of "mapping harvesters" to handle known IP address mappings. A set of mapping harvesters is configured once when the library initializes, and each of them contains a pair of IP addresses (local and public).

When an ICE Agent gathers candidates, it uses the set of mapping harvesters to obtain srflx candidates without the use to e.g. a STUN server dynamically.

Mapping harvesters preserve the port number of the original candidate, so they should only be used when port numbers are preserved.

Ice4j implements three types of mapping harvesters: one with a pre-configured pair of addresses, one two which discover addresses dynamically using the AWS API and STUN.

org.ice4j.ice.harvest.NAT_HARVESTER_LOCAL_ADDRESS

org.ice4j.ice.harvest.NAT_HARVESTER_PUBLIC_ADDRESS

Default: none

Configures the addresses of the pre-configured mapping harvester.

org.ice4j.ice.harvest.DISABLE_AWS_HARVESTER

Default: false

Explicitly disables the AWS mapping harvester. By default the harvester is enabled if ice4j detects that it is running in the AWS network.

org.ice4j.ice.harvest.FORCE_AWS_HARVESTER

Default: false

Force the use of the AWS mapping harvester, even if ice4j did not detect that it is running in the AWS network.

org.ice4j.ice.harvest.STUN_MAPPING_HARVESTER_ADDRESSES

Default: none

A comma-separated list of STUN server addresses to use for mapping harvesters. Each STUN server address is an ip_address:port pair. Example: stun1.example.com:12345,stun2.example.com:23456

Single port UDP harvester

org.ice4j.ice.harvest.AbstractUdpHarvester.SO_RCVBUF

Configures the receive buffer size for the single port UDP harvester (or other AbstractUdpListener implementations). If this is not set the system default value will be used (the net.core.rmem_default sysctl parameter on linux). Note that the system maximum value may need to be increased (net.core.rmem_max on linux).