Multiplayer K: events flood

[JardY]

Hi A2A,

First, congratulations for your 172. Hard work to great great result !

Sorry for my english, i'm french so i'll try to explain my problem with the good words.

We host a session (http://www.praetorians-fs.com), and we have a monitoring system which monitor events (K: events in xml programming). I don't want to learn you this. I'm sure that the A2A dev team knows this but almost of the K: events sent an network information (UDP packet). When a session is flooded by events, players/pilots experiment desynchronizations with the placement of other aircrafts or outright disconnections.

Come to the point : In the gauge code_c172r.xml, you use SPOILERS_SET event (I guess this is to simulate the drag). The code is pretty complex and we can reduce the gauge frequency from 18 Hz to 1 Hz. So, it can go, but there is another flood with SMOKE_ON & SMOKE_OFF (perhaps, in DLL...), can you try to check the state of the smoke system before trigger the K:SMOKE_ON & K:SMOKE_OFF. This will drastically reduce the "network" impact of your aircraft.

Do not think I'm here to teach you how to develop an aircraft, but just to bring you our experience about network & multiplayer FSX session.

Feel free to ask me for any extra informations.

Thanks for reading me and thanks again for your great work !

JardY.

[N6722c]

Hi A2A,

First, congratulations for your 172. Hard work to great great result !

Sorry for my english, i'm french so i'll try to explain my problem with the good words.

We host a session (http://www.praetorians-fs.com), and we have a monitoring system which monitor events (K: events in xml programming). I don't want to learn you this. I'm sure that the A2A dev team knows this but almost of the K: events sent an network information (UDP packet). When a session is flooded by events, players/pilots experiment desynchronizations with the placement of other aircrafts or outright disconnections.

Come to the point : In the gauge code_c172r.xml, you use SPOILERS_SET event (I guess this is to simulate the drag). The code is pretty complex and we can reduce the gauge frequency from 18 Hz to 1 Hz. So, it can go, but there is another flood with SMOKE_ON & SMOKE_OFF (perhaps, in DLL...), can you try to check the state of the smoke system before trigger the K:SMOKE_ON & K:SMOKE_OFF. This will drastically reduce the "network" impact of your aircraft.

Do not think I'm here to teach you how to develop an aircraft, but just to bring you our experience about network & multiplayer FSX session.

Feel free to ask me for any extra informations.

Thanks for reading me and thanks again for your great work !

JardY.

Confirmed:


In the gauge code_c172r.xml, SPOILERS_SET event code looks "Suspect" in case.
You might want to look at the generation of L:AOA_Limiter1 and L:AOA_Limiter2
Is this really for a C172 ???
In any case, the value that you are setting the Spoilers to is always 0, but it is being set at Gauge rate, (18 / second) when plane is in the air.
If you check to see (A:SPOILERS HANDLE POSITION, position) first, and only call the key event if the value is incorrect, you will fix the Spoiler flooding.

In particular check generation of L:AOA_Limiter1 ( always zero, as multiplied by 0 )

Since anyone can look at this XML, here is a suggested correction.

LOGIC

Code: Select all

if ( (A:SIM ON GROUND , bool) == 0 ) 
 {	 
 	 if ( (A:INCIDENCE ALPHA, degrees) > 45 ) 
 	 	{
		(L:AOA_Limiter1, number) = max( ( ( ( (A:INCIDENCE ALPHA, degrees) - 28 ) * ( ( abs( (A:VELOCITY BODY Y, feet per minute) ) - 1000 ) / 500 ) ) * 0 ),0 ) ;    // always zero  ???
		  
		(L:AOA_Limiter2, number) = max( ( ( ( (A:INCIDENCE ALPHA, degrees) - 45 ) * ( ( abs( (A:VELOCITY BODY Y, feet per minute) ) - 1000 ) / 500 ) ) * 0.25 ),0 ) ;  //  ?????? 
		  
		 	 
		 (L:AOA_Limiter, number) =  (L:AOA_Limiter1, number) + (L:AOA_Limiter2, number);  
		  
 		 if ( (A:SPOILERS HANDLE POSITION, position) != (L:AOL_Limiter, number) ) 
 		 	{
			 (K:SPOILERS_SET) = (L:AOA_Limiter, number) ;    
		 	} 
		 }
		 	
} else  {
		(L:AOA_Limiter, number) = 0 ;  
		 (L:SpoilerINI, bool) = 0 ;  
		  
 		if ( (A:SPOILERS HANDLE POSITION, position) != 0 ) 
 		 	{
			 (K:SPOILERS_SET) = 0 ;  
		 	} 
} 

XML

Code: Select all

(A:SIM ON GROUND , bool) 0 == if{ (A:INCIDENCE ALPHA , degrees) 45 > if{ (A:INCIDENCE ALPHA , degrees) 28 - (A:VELOCITY BODY Y , feet per minute) abs 1000 - 500 / * 0 * 0 max (>L:AOA_Limiter1 , number) (A:INCIDENCE ALPHA , degrees) 45 - (A:VELOCITY BODY Y , feet per minute) abs 1000 - 500 / * 0.25 * 0 max (>L:AOA_Limiter2 , number) (L:AOA_Limiter1 , number) (L:AOA_Limiter2 , number) + (>L:AOA_Limiter , number) (A:SPOILERS HANDLE POSITION , position) (L:AOL_Limiter , number) != if{ (L:AOA_Limiter , number) (>K:SPOILERS_SET) } } } els{ 0 (>L:AOA_Limiter , number) 0 (>L:SpoilerINI , bool) (A:SPOILERS HANDLE POSITION , position) 0 != if{ 0 (>K:SPOILERS_SET) } } 

TESTED: No Spoiler Flooding.

Since I do not really understand what you were intending to do here, I may be completely off base, but even if so, this is where you might want to look for a solution.

[Gypsy Baron]

Hello JardY,

I called attention to this thread the team to look at. Since I fly Multiplayer
almost exclusively, I usually test and look for K:events during testing that
might flood a MP server, however I have been traveling a great deal all summer
and did not get a chance to look at this aspect of the C172 before release.
I last looked back in June and didn't see any flooding at that point so some
of the changes and evolution since that time may have introduced the events
you are seeing.

Rest assured that the team is aware of the MP environment requirements
with regards to server flooding. I'm sorry this one 'got by me'.

Paul

[JardY]

First, thanks a lot for your answer !

As N6722c said, instead of reduce the frequency of gauge code_c172r, protect the SPOILERS_SET calls by checking the position of the spoilers.

Will try this.

No worry and take your time ! The 777 PMDG just released and some of my pilots told me 777 flooding... I'll be MORE MORE MORE difficult to explain it at PMDG

Regards and thanks again

JardY.

[Lewis - A2A]

Hey guys,

thanks for the report. We had done some limited MP testing and thought we had got all the flooding. The team will look into it

thanks,
Lewis - A2A

[JardY]

Thanks again.

The log of the SMOKE_ON when the plane is in cold & dark state :

(Each 20 seconds, our monitoring system logs the events and how many times one kind of event was received)

Even my English is very... bad, I can send and explain to A2A how to connect to our session and how to follow the monitoring log like below :

20130904_132855|3B:03|TOGGLE_ALTERNATOR1|1
20130904_132855|0D:02|LANDING_LIGHTS_SET|3
20130904_132855|BF:02|TOGGLE_BEACON_LIGHTS|1
20130904_132855|06:02|STROBES_SET|3
20130904_132855|8D:04|AVIONICS_MASTER_SET|4
20130904_132855|01:02|SMOKE_ON|13
20130904_132855|02:02|SMOKE_OFF|1
20130904_132855|20:02|GEAR_DOWN|1
20130904_132905|01:02|SMOKE_ON|37
20130904_132915|01:02|SMOKE_ON|35
20130904_132925|01:02|SMOKE_ON|36
20130904_132935|01:02|SMOKE_ON|37
20130904_132946|01:02|SMOKE_ON|37
20130904_132956|01:02|SMOKE_ON|36
20130904_133006|01:02|SMOKE_ON|36
20130904_133016|01:02|SMOKE_ON|36
20130904_133016|00:DF|UNKNOWN|1
20130904_133026|01:02|SMOKE_ON|37
20130904_133036|01:02|SMOKE_ON|36
20130904_133036|20:02|GEAR_DOWN|1
20130904_133046|01:02|SMOKE_ON|36
20130904_133057|01:02|SMOKE_ON|38
20130904_133107|01:02|SMOKE_ON|36
20130904_133117|01:02|SMOKE_ON|36
20130904_133127|01:02|SMOKE_ON|36
20130904_133137|01:02|SMOKE_ON|38
20130904_133147|01:02|SMOKE_ON|36
20130904_133157|01:02|SMOKE_ON|36
20130904_133208|01:02|SMOKE_ON|36
20130904_133218|01:02|SMOKE_ON|38
20130904_133218|20:02|GEAR_DOWN|1
20130904_133228|01:02|SMOKE_ON|36
20130904_133238|01:02|SMOKE_ON|35
20130904_133238|00:DF|UNKNOWN|1
20130904_133248|01:02|SMOKE_ON|36
20130904_133259|01:02|SMOKE_ON|40
20130904_133309|01:02|SMOKE_ON|38

Good luck and thanks again.

Edit : When the engine is started, SMOKE_ON stops.

Edit 2 : Even if the plane is airborn the SMOKE_ON & SMOKE_OFF are triggered something... Just try to check the SMOKE SYSTEM state before to correct the flood.

[N6722c]

The 777 PMDG just released and some of my pilots told me 777 flooding... I'll be MORE MORE MORE difficult to explain it at PMDG

Try IMPOSSIBLE !!

[JardY]

The LANDING_LIGHTS_ON flood will be fix next SP

In french : "Victoire !!!"

[ak416]

Just wanted to say thanks to everyone involved for looking into this. A lot of us in the group I fly with are big A2A fans (including Captain Jakey ) and multiplayer is essentially all that we do. and fixing seemingly small issues like these go a long way to increase server stability and provide a better experience for everyone involved.