I challenge Frontier Developements!

And it is also why many posters - who I suspect of being aligned with the developers - say they don't want features or game aspects that are resource heavy on the P2P client instancing model...

Makes perfect sense.

63984465.jpg
 
If only ISP's actually delivered the linerates they promised and got rid of datacaps, traffic shaping, throttling, preferred peerage. If only the comms companies actually invested in infrastructure. If only every internet-connected device was on a secure fibre channel. If only we could change the speed of light.

Until then - you can absolutely bet every penny you will ever earn that every man, his dog, and it's fleas will mess with their connections to gain some benefit over a competitor.

Yes, the fiber for everyone
 
Yes, the fiber for everyone

Still won't help.

LINK - Other side of the world to FDs servers on a single cable using optimal conditions and ignoring speed of light slowdown in a fibre medium ~133ms.

Realistically at 3,500KM away from FD my ping is around 150-200ms - so the OP is either suggesting somehow FD change the laws of physics or they exclude some people from the game as there's no amount of optimisation that can fix that*



EDIT:
*I am aware FD try to group people together from similar regions to help reduce latency for the P2P links which helps.
 
Last edited:
How do you check your ping to frontiers servers?

+I thought the multiplayer stuff was all p2p, and servers were just for looking up basic system data etc. So presumably the servers wouldn't need a low ping. All that should matter is your own net connection and that of whoever else is in your instance. Or do I have it wrong?
 
p2p design is up to fdev infinite greed and bobbleheads. P2p design is wrong for real time multiplayer and this has been known for many years.

Er by who? The wise goblins of the interwebs?

There are quite a few games using P2P and distributed systems are considered the way forward in computer science.

If you want to make a dedicated server multiplayer game you are limiting your game design by the number of possible instances of player connections. That usually means arena games or a smaller number of dedicated locations holding known instances. I can't see Elite being Elite if it was designed like that, people would complain that the old Elite games had the entire galaxy to explore bla bla. There would be no exploring the center of the galaxy with your friends, no distant worlds - to do that would require dedicated server instances for every possible combination of player groupings that went of in any directions. Name a game that provides that with servers?

Sure, you can spin up shards in the cloud on demand but that doesn't happen instantly and the costs would be ridiculous. Monthly fee wouldn't cover it. Again, if you think it can be done, give examples - always interested to read about new tech.

Then you have the problem of real-time simulation games that throw around a huge amount of network code and the limitations on network technology to deal with it. There was an Elite inspired multiplayer game that tackled these issues before and ended up becoming EvE, point & click, command & control - just like most multiplayer games that want to have large scale player interactions. If Elite did that it wouldn't be an Elite game, it would just be another EvE, no cockpits, no wizzing past other ships at close range gripping the joystick, no virtual-reality flight experience.

In making a new Elite game I imagine Frontier had 3 choices, a dedicated server game that would look absolutely nothing like any previous Elite game, removing multi-player altogether (would please some people here, but nothing to do with this thread), or a distributed network.

They hybrid cloud system they developed isn't as cheap as people seem to think, I can't see how "greed" would have inspired that design when it's very obvious to me that Elite inherently wouldn't work without such a system.
 
Last edited:
If games like Battlefield 3/4 can handle 64 players with all bullets and environmemt destruction at an average of 100ms then ED should be able to handle two players in one instance with 100ms as well but not 2000. There is no slider.

It simoky is possible to play with more players together at a decent ping rate. The slider is more like a mathmatical curve and expanding the curve means (economical) growth. Frintiers muktiplayer slide is pretty small tho at this moment :(

You can't really compare these two games. I'm sure if FDEV had the revenue of an entity like EA then yes, 64 player low latency (though debatable even then) instances based on thousands of globally located servers would be great.

I really don't see FDEV having the player base or the monthly revenue a company like EA has. (I think it should though :) )

Correct me if I'm wrong by all means but when did EA have to Kickstart a game?
 
Still won't help.

LINK - Other side of the world to FDs servers on a single cable using optimal conditions and ignoring speed of light slowdown in a fibre medium ~133ms.

Realistically at 3,500KM away from FD my ping is around 150-200ms - so the OP is either suggesting somehow FD change the laws of physics or they exclude some people from the game as there's no amount of optimisation that can fix that*



EDIT:
*I am aware FD try to group people together from similar regions to help reduce latency for the P2P links which helps.



I think you will find the OP is not asking Fdev to change to laws of physics....

They just want the game to work correctly, according to what is written on the tin, and not have the multiplayer aspect of the game occasionally sux....



The background networking and cheap P2P client instancing Fdev have used to make Elite:Dangerous do not lend themselves to a good MASSIVE MULTIPLAYER experience, because it is a multiplayer game with an already high P2P network package volume, due to the freedom of "non-targeted" "non-time-based actions". This package volume and bandwidth use of the P2P client instancing exponentially increases as players join an instance - but Fdev have gone ahead any way and chosen to make a solo, or at best a co-op game into an MMO.

Potentially one of the reasons Elite: Dangerous is also failing as the OP states to be a successful co-op experience is because so few people are playing globally - and the network is forced to join people together from larger global distances - impacting customers who do not play in a global region close to any large cluster of players.

Truly it is a) it is a niche game, and b) it has been getting a bad reception from customers in the last 4-5 months, as non-kickstarters buy the game.


Lets face it - making a true MMO out of Elite: Dangerous is almost impossible to achieve unless you have dedicated global server shards - for which a subscription pay system would be needed to support. Alternatively you can drastically trim back standard MMO game features that use up even the smallest amount of bandwidth.

And we know which route Fdev have taken - not wanting a subscription base format they were forced to use P2P client instancing. So - to reduce the bandwidth load on the instancing, and background simulation they have omitted or minimized many standard MMO features seen in current and historical MMOs such as global chat, guilds AND player resource persistent activities like P2P trading, auction houses, station module or trade item storage and ship customization.


SUMMARY

To be fair - in my opinion this MMO "P2P client instancing" is not fit for purpose for a game that has such great ambitions as Elite: Dangerous does - and I am sure we will see similar reactions to it's contemporaries once they claim their products are also no longer in beta...
 
It's quite clear that you have no real understanding of networking.

Please purchase a copy of the TCP/IP Illustrated volumes and read them. Read them again. Then you will understand.

No he won't.

You have any idea how many full-time IT pros I've encountered over the last 30 years that can't get their head around TCP/IP and the 7-layer model?
 
No he won't.

You have any idea how many full-time IT pros I've encountered over the last 30 years that can't get their head around TCP/IP and the 7-layer model?
That is because no-one actually implements all 7 layers, its 5 at best in reality. But its a fun trivia question for when you interview people anyway.
 
Still won't help.

LINK - Other side of the world to FDs servers on a single cable using optimal conditions and ignoring speed of light slowdown in a fibre medium ~133ms.

Realistically at 3,500KM away from FD my ping is around 150-200ms - so the OP is either suggesting somehow FD change the laws of physics or they exclude some people from the game as there's no amount of optimisation that can fix that*



EDIT:
*I am aware FD try to group people together from similar regions to help reduce latency for the P2P links which helps.


It seems people don't read carefully.
I have challenged them to provide a ping of 100ms for LOCAL players (UK - UK for example) for (US - UK) I am fine with 200-300. But a 2000ms, no matter the location is not breaking the laws but an incredibly low-end networking architecture.
Yes it coests money but so does ED and Horizons no? Aren't we allowed to expect certain basics for our cash?
 
No he won't.

You have any idea how many full-time IT pros I've encountered over the last 30 years that can't get their head around TCP/IP and the 7-layer model?

That's because knowing the "7-layer network model" is irrelevant and pointless for almost everybody in IT. :)
 
Back
Top Bottom