On 02/23/2016 02:56 PM, Gordon Sim wrote:
On 23/02/16 18:53, Ted Ross wrote:Here's what I'm using in the tross-DISPATCH_179-1 development branch. Like you suggest, this combines multiple legacy entities into a single entity. In this case, the new entity type replaces "fixedAddress", "linkRoutePattern", and "waypoint". Fixed address and link route pattern will remain supported in the config file for backward compatibility. It's likely that waypoint will also remain backward compatible, but I'm not sure yet. I didn't attempt to provide a meaningful name like "relay" and just called it "provisioned" as it will the primary entity type for provisioning the wiring inside the router. "router.provisioned" { "description": "Provisioned component within the router: Address configuration, waypoint, or link-route destination" "attributes": { "objectType": {"type": ["address", "linkDestination", "waypoint"]} "address": {}, "connector": {}, "direction": {}, "semantics": {}, "ingressAddress": {}, "egressAddress": {}, "ingressSemantics": {}, "egressSemantics": {} }I like both your and Alan's attempts to simplify/unify the configuration. I do think 'provisioned' is too opaque a name though. What about 'route' with types 'simple', 'link' or 'path'? If 'ingressAddress' and 'egressAddress' are specified, can 'address' also be specified and if so what would it mean?
For waypoints, {ingress,egress}Address are the phase-annotated versions of 'address', which is the externally-visible address.
Affect on waypoint configuration: The legacy way to configure a simple waypoint is: waypoint { address: my_queue connector: my_broker inPhase: 0 outPhase: 1 } fixedAddress { prefix: my_queue phase: 0 fanout: single bias: spread } fixedAddress { prefix: my_queue phase: 1 fanout: single bias: spread } The new way is: provisioned { objectType: waypoint address: my_queue connector: my_broker }Does that mean that 'single fanout with spread' semantics are assumed by default for any waypoint? (I think the default fanout mode and bias should be the same regardless of the 'type').
Yes, and yes.
--------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
--------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
