The Battle of Asakai: How EVE Online handles giant battles
The battle of Asakai was the second in the number of ships encountered in the history of EVE Online ( Komrad madmaxcorp wrote this beautifully, in the same place you can watch the eyewitness videos without exaggerating the galactic beauty). The battle history is interesting in itself (human error), but traditionally we are more interested in what happened to the game servers during this crazy battle, how did CCP (creator of EVE Online) manage to cope with such loads?
“Our support team (game masters, GMs) follow giant battles like this,” says one developer. “We have a web page with the status of a cluster, on which large red numbers appear in case of overloading of nodes in battles. So it's pretty easy to see what's going on. ” More interesting is what is behind these numbers.
')
The most common answer is to do nothing. Almost all battles create a sane load on the server, or they are so fast that nothing can be done. However, as soon as it becomes clear that doing something is necessary, the team has several opportunities.
It is difficult to notice, but sometimes admins change the game space itself. “We are moving other solar systems from the node on which the battle is fought. This temporarily disables everyone who is in these systems, but relieves them of the symptoms inherent in an overloaded server. ” - explains the developer. “It helps a little to the system in which the battle takes place, especially if the fleets of reinforcements pass through portable systems. This was done in the case of the last battle, but this rarely happens. "
There are more significant weapons to support the performance of servers under extreme load, but they are already compromise. "We have a pair of machines with the best iron, better than the average in the cluster."
“These machines come into play when players need it. Unfortunately, the same thing happens here that was described above - everyone in the system is disconnected during the transfer. Because of this, we basically never use such a move, ”he continues. “In this case, the battle will simply fall apart, since Titan will take a random position. If we had gone this way and moved the system, Titan and his friends would simply not have returned, interrupting the battle. So yes, we did not. ”
So, solutions to the problem are actually not enough. Servers are always 100% loaded, so nothing can be tweaked in the settings. However, the game has a built-in mechanism for solving the problem of big battles. She slows down.
This has long been a familiar feature for EVE Online veterans, but a curious way of dealing with high loads and lags in a situation where the game should work. At that moment, when the load on the server reaches a critical point, the game automatically slows down time by a certain amount to counter the overload. The time in the battle for 3000 players is running at 10% of the real, which is the maximum possible slowdown. Here is a video that demonstrates the time dilation in practice.
“This is a very elegant way to cope with lags in a situation where the servers of other games simply melt,” the author writes. “In fact, time in the system is slowed down so that requests to the server and its responses are implemented and implemented correctly. In this case, people entering the game get gradually slowing down to a limit of 10% of the time - in fact, the world is slow, but quite workable in terms of gameplay and tactics. ”
So instead of dropping the servers, time dilation allows even massive battles to be realized, albeit at a slow pace. As a result, there may be battles like this one, in which thousands of players participate with thousands of expensive ships.
“For the HoneyBadgers clan, the losses were 6 Dreadnoughts, 11 Aircraft Carriers and one Supercarrier. For Clusters, everything turned out worse - 4 Dreadnought, 29 Aircraft Carriers, 5 Supercarriers and three Titans, ”writes PC Gamer. The battle of Asakai is called a complete rout.
Approximately losses amounted to 700 billion in game currency, or $ 24,000. And that is why it is so important that such battles take place on stable servers. Fortunately, CCP was ready for such an event. And EVE Online continues to be an exciting game.
“Our support team (game masters, GMs) follow giant battles like this,” says one developer. “We have a web page with the status of a cluster, on which large red numbers appear in case of overloading of nodes in battles. So it's pretty easy to see what's going on. ” More interesting is what is behind these numbers.
')
Available options
The most common answer is to do nothing. Almost all battles create a sane load on the server, or they are so fast that nothing can be done. However, as soon as it becomes clear that doing something is necessary, the team has several opportunities.
It is difficult to notice, but sometimes admins change the game space itself. “We are moving other solar systems from the node on which the battle is fought. This temporarily disables everyone who is in these systems, but relieves them of the symptoms inherent in an overloaded server. ” - explains the developer. “It helps a little to the system in which the battle takes place, especially if the fleets of reinforcements pass through portable systems. This was done in the case of the last battle, but this rarely happens. "
There are more significant weapons to support the performance of servers under extreme load, but they are already compromise. "We have a pair of machines with the best iron, better than the average in the cluster."
“These machines come into play when players need it. Unfortunately, the same thing happens here that was described above - everyone in the system is disconnected during the transfer. Because of this, we basically never use such a move, ”he continues. “In this case, the battle will simply fall apart, since Titan will take a random position. If we had gone this way and moved the system, Titan and his friends would simply not have returned, interrupting the battle. So yes, we did not. ”
So, solutions to the problem are actually not enough. Servers are always 100% loaded, so nothing can be tweaked in the settings. However, the game has a built-in mechanism for solving the problem of big battles. She slows down.
Time slowdown
This has long been a familiar feature for EVE Online veterans, but a curious way of dealing with high loads and lags in a situation where the game should work. At that moment, when the load on the server reaches a critical point, the game automatically slows down time by a certain amount to counter the overload. The time in the battle for 3000 players is running at 10% of the real, which is the maximum possible slowdown. Here is a video that demonstrates the time dilation in practice.
“This is a very elegant way to cope with lags in a situation where the servers of other games simply melt,” the author writes. “In fact, time in the system is slowed down so that requests to the server and its responses are implemented and implemented correctly. In this case, people entering the game get gradually slowing down to a limit of 10% of the time - in fact, the world is slow, but quite workable in terms of gameplay and tactics. ”
So instead of dropping the servers, time dilation allows even massive battles to be realized, albeit at a slow pace. As a result, there may be battles like this one, in which thousands of players participate with thousands of expensive ships.
“For the HoneyBadgers clan, the losses were 6 Dreadnoughts, 11 Aircraft Carriers and one Supercarrier. For Clusters, everything turned out worse - 4 Dreadnought, 29 Aircraft Carriers, 5 Supercarriers and three Titans, ”writes PC Gamer. The battle of Asakai is called a complete rout.
Approximately losses amounted to 700 billion in game currency, or $ 24,000. And that is why it is so important that such battles take place on stable servers. Fortunately, CCP was ready for such an event. And EVE Online continues to be an exciting game.
Source: https://habr.com/ru/post/167651/
All Articles