Jump to content
Sign in to follow this  
You need to play a total of 10 battles to post in this section.
Dardanelles

Why cant WOW tell the difference between non participation and getting dumped from the server?

22 comments in this topic

Recommended Posts

Member
1 post
361 battles

I returned to the game a couple of days ago and had a great couple of evenings gaming but then this morning had the server / connection lag and drop out on multiple occasions, making the game unplayable.  I logged off and then when I logged back on I wind a warning for conduct violation and then I get logged off the game again.

Now I'm not sure whether to try playing until this issue is resolved as I'm already on a warning for issues beyond my control and if it happens again, i.e. lag or get dumped from the server, then while I'm desperately trying to login before I'm sunk, someone will have a whinge about me not participating and I'll get banned.  

Share this post


Link to post
Share on other sites
[CLAY]
Member
473 posts
6,471 battles
12 hours ago, RalphTheTheatreCat said:

The Reason the game doesn't differentiate because there is nothing stopping a player just switching off the modem to fake a disconnect. (Just one of a out of scenarios).

So in summary legit players get punished because devious players exploit the system.

Unfair, but until we force the bad ones to behave, it’s something we have to deal with.

Not that getting pink is that big of a deal. A few co-op games and you’re right. If you keep disconnecting and getting pink, may be time to call it a night and hope for a better connection tomorrow.

Edited by Grygus_Triss

Share this post


Link to post
Share on other sites
Super Tester
4,012 posts
4,448 battles

Its a game engine...…..not NASA.  It can only do so much.  Gaming is based on the lowest common denominator (as is life) We don't make rules to accommodate the people who do the right thing.  We develop them to control the detritus of society.  Games are a microcosm of real life in that people will exploit it to their advantage and remain unchecked until sanctioned.

Share this post


Link to post
Share on other sites
Member
134 posts
3,425 battles
8 minutes ago, RalphTheTheatreCat said:

Games are a microcosm of real life in that people will exploit it to their advantage and remain unchecked until sanctioned

but IRL the gun and bullet work very well :Smile_hiding:

Share this post


Link to post
Share on other sites
Super Tester
7,897 posts
8,143 battles

Now you know the game can't. So I think next time you get dropped you won't be surprised.

Also, getting non participation penalty from getting dumped from server is really hard. If you could not log back in within 5 minutes, no matter what happened, from the perspective of other 11 players, you did not participate. And you really need to blame your internet for that. If server dumps you, it dumped all players as well, the matches are not counted for those rare cases and no one gets penalty.

So yah, "getting dumped from the server" kinda gives the wrong notion.

Share this post


Link to post
Share on other sites
Member
100 posts
1,446 battles

Switch Off/Unplugging can be detected by OS,  sand the status can be retrieved by Bigworld Engine which WOWS built on.  So does bad and intermediate connections.

The game monitoring mechanism even scan the memory for aim bot at the same time tracking crash and generating report in real time. 

It is Totally Not True the game can not differentiate crashes and problem from deliberate tampering with the game. WG choose to use them for finding user side problems only.

Share this post


Link to post
Share on other sites
Super Tester
7,897 posts
8,143 battles
29 minutes ago, tsuenwan said:

Switch Off/Unplugging can be detected by OS,

And I'm sure you know how OS could do that, right? Cause from what I studied at universities, it should not be able to. As I understand, OS can only tell there is no connection, it cannot tell "Why" if that's an external incident.

Edited by icy_phoenix

Share this post


Link to post
Share on other sites
Member
100 posts
1,446 battles
41 minutes ago, icy_phoenix said:

And I'm sure you know how OS could do that, right? Cause from what I studied at universities, it should not be able to. As I understand, OS can only tell there is no connection, it cannot tell "Why" if that's an external incident.

It depends on what you actually studied in universities.

Unplugged cable and switch modem would had no power can be detected by network interface, Unplugged internet connection between ISP and modem would had no internet connection which can be detect by the network stack. connection performance can be tracked by whole bunch of network diagnostic functions available with the OS. 

Ultimately, BigWorld engine tracks data transfer (sucessful or lost) rate between the client and server.

Distributed computing is old, Long before those venture capitalist renamed it to become "Cloud XY&*BBBSSS". There were many means developed there to track where the problem lays. Only if you want to use them.

In fact WG support knew all this well. but only use them when situation at their convenience.

I have worked long enough to see Software people would always blamed the user, the network and the OS rather then software problem. and Network people would conveniently blame the user, the software and the OS rather then the network connections. 

Edited by tsuenwan

Share this post


Link to post
Share on other sites
Member
100 posts
1,446 battles

 

When your network equipment or ISP was blamed for network problem. there are more than one internet connection test website around the world to test you internet connection and performance to many countries including connection to Japan. ie: http://www.myspeedmeter.net/

Share this post


Link to post
Share on other sites
Super Tester
7,897 posts
8,143 battles
4 hours ago, tsuenwan said:

It depends on what you actually studied in universities.

Unplugged cable and switch modem would had no power can be detected by network interface, Unplugged internet connection between ISP and modem would had no internet connection which can be detect by the network stack. connection performance can be tracked by whole bunch of network diagnostic functions available with the OS. 

Ultimately, BigWorld engine tracks data transfer (sucessful or lost) rate between the client and server.

Distributed computing is old, Long before those venture capitalist renamed it to become "Cloud XY&*BBBSSS". There were many means developed there to track where the problem lays. Only if you want to use them.

In fact WG support knew all this well. but only use them when situation at their convenience.

I have worked long enough to see Software people would always blamed the user, the network and the OS rather then software problem. and Network people would conveniently blame the user, the software and the OS rather then the network connections. 

I get your point about software people often blaming others than softwares, that happens, not denying. But this wasn't the question I raised. I asked how OS could detect whether a NIC power was cut off or a thunderbolt struck the line in between your pc and the next point. Both results in a no-connection signal to be raised. I am asking again. cause I am genuinely interested as computer science was my major. All im getting from your post is (a) some hunch about OS, (b) slight misconception about a game engine, (c) rant against capitalism, (d) little support bashing, (e) blaming software people. Let's talk technical terms, shall we?

Anyway here is something that I'm quoting from a credible source:

Quote

 

Causes of Half-Open Connections

Half-open connections are in that annoying list of problems that one seldomly sees in a test environment but commonly happen in the real world. This is because if the socket is shut down with the normal four-way handshake (or even if it is abruptly closed), the half-open problem will not occur. Some of the common causes of a half-open connection are described below:

  • Process crash. If a process shuts down normally, it usually sends out a “FIN” packet, which informs the other side that the connection has been lost. However, if a process crashes or is terminated (e.g., from Task Manager), this is not guaranteed. It is possible that the OS will send out a “FIN” packet on behalf of a crashed process; however, this is up to the OS.
  • Computer crash. If the entire computer (including the OS) crashes or loses power, then there is certainly no notification to the other side that the connection has been lost.
  • Router crash/reboot. Any of the routers along the route from one side to the other may also crash or be rebooted; this causes a loss of connection if data is being sent at that time. If no data is being sent at that exact time, then the connection is not lost.
  • Network cable unplugged. Any network cables unplugged along the route from one side to the other will cause a loss of connection without any notification. This is similar to the router case; if there is no data being transferred, then the connection is not actually lost. However, computers usually will detect if their specific network cable is unplugged and may notify their local sockets that the network was lost (the remote side will not be notified).
  • Wireless devices (including laptops) moving out of range. A wireless device that moves out of its access point range will lose its connection. This is an often-overlooked but increasingly common situation.

In all of the situations above, it is possible that one side may be aware of the loss of connection, while the other side is not.

 

As it says, for all these cases, the client end may not know what actually happened. I'm sure it's just an attempt from software people to blame something else, but I hope it proves to be helpful.

 

Also, I should probably post about what ping actually is and what those speed testers actually do, but I'm kinda lazy.

Edited by icy_phoenix

Share this post


Link to post
Share on other sites
Member
211 posts
11,973 battles

You're playing a video games and you want its server or the providers to gathers info on the reason for inactivity !?
All it needs to do is find out where you live and see if there's electricity problem, provider problem, etc. !?

Errr sorry ? Am I high ? Because all that sounds like you want someone to gather your privacy informations through a video game.
Do you happens to be living in the dark web all this time and just came out of the dark now?
I'm not high right !?

Share this post


Link to post
Share on other sites
Member
100 posts
1,446 battles
13 hours ago, icy_phoenix said:

I get your point about software people often blaming others than softwares, that happens, not denying. But this wasn't the question I raised. I asked how OS could detect whether a NIC power was cut off or a thunderbolt struck the line in between your pc and the next point. Both results in a no-connection signal to be raised. I am asking again. cause I am genuinely interested as computer science was my major. All im getting from your post is (a) some hunch about OS, (b) slight misconception about a game engine, (c) rant against capitalism, (d) little support bashing, (e) blaming software people. Let's talk technical terms, shall we?

Anyway here is something that I'm quoting from a credible source:

As it says, for all these cases, the client end may not know what actually happened. I'm sure it's just an attempt from software people to blame something else, but I hope it proves to be helpful.

 

Also, I should probably post about what ping actually is and what those speed testers actually do, but I'm kinda lazy.

When people talking about faking crashes by removing there Local cables, switch off their modems etc. These are fully dropped connections,,  and There are many means in the OS's network stack to reflect the power status of the cable.

And, when talking about next point behind the ISP. and how to locate the problem along the data path in the so called half open connection,  then there are network tracing functions can be used to find out how far your routing could reach out.  Please also note that internet was developed based on packet switching but circuit switch technologies. and data path between ISPs and Telecoms are more likely have multiple physical redundancies.

and talking about how. in the network stack NIC operation state reflects cable plug in and power states. carrier states reflects if data are present. there are programming API as well as commands such ifconfig/ipconfig shows them

Where are connection problem along the data path, long before the connection actually dropped. network packet may be corrupted or dropped, Data timeout and error will increase. and these Data transfer error rate can be read by netstat down to per interface per port level.  and how data path is routed in Hops can be traced by traceroute.  and once you know how your traffic is routed to the destination. you may use ping to ping not just the destination but also the routing point along you data path.

In the windows event viewer, cable unplugged between modem/pc will reflect as disconnection,  and cable unplug between modem to ISP will reflect as local network only or no internet.  when you connected to the ISP, you will get an different assigned IP then you can start looking at the data path.

Talking about game engine concept. MMOG engine would need to allow the game to work with a combination of connections with very different network performance. Earlier I used Wireshark to look at what WGCheck actually did. and it did volume transfer test with the server apart from Ping. which are also found in Unity and Unreal etc engines.

Internet performance test websites allows user to test network performance between countries, and it is the simplest way to test the internet path and its performances.

 

Edited by tsuenwan

Share this post


Link to post
Share on other sites
Moderator
592 posts
2,995 battles
20 hours ago, icy_phoenix said:

And I'm sure you know how OS could do that, right? Cause from what I studied at universities, it should not be able to. As I understand, OS can only tell there is no connection, it cannot tell "Why" if that's an external incident.

Hehe I pop back in for a second and I see this 😛

*puts on my MCSE hat (Microsoft Certified Systems Engineer)*

Just off the top of my head - Windows since 7 does record a comprehensive set of codes and event ID's in order to determine why it's restarted. 

The three below are the ones I know off by heart, and get recorded at time of Shut Down, so in theory, they could be sent up to Wargaming as a better indication as to why the player has suddenly dropped. 

Am I saying that this occurs, no, well it depends, I've never acutally looked at the code in WG games, nor do I care about it that much. 

BUT 🙂 Yes, windows does track states and why it was shut down. 

Additionally - 
 

Event ID Description
1074 Logged when an app (ex: Windows Update) causes the system to restart, or when a user initiates a restart or shutdown.
6006 Logged as a clean shutdown. It gives the message "The Event log service was stopped".
6008 Logged as a dirty shutdown. It gives the message "The previous system shutdown at time on date was unexpected".

 

In addition we have the additional logs that can determine action done by a user. 

4647 - Log off and Sign Out

For the life of me, i also can't remember the event ID number that tells you the Network State as well, not to mention battery health and sleep modes for Windows. But they are out there and if i worried more about it i would search. 

So as for your statement, Windows 100% knows if it was a clean power down / log off / application or a dirty shutdown. What It doesn't know is the 'exact' reason why and if you know the server family of windows, that's why when you boot back in, it will always prompt you for a 'why' and what happened. 

 

Share this post


Link to post
Share on other sites
Super Tester
7,897 posts
8,143 battles
1 hour ago, tc1259 said:

Hehe I pop back in for a second and I see this 😛

*puts on my MCSE hat (Microsoft Certified Systems Engineer)*

Just off the top of my head - Windows since 7 does record a comprehensive set of codes and event ID's in order to determine why it's restarted. 

The three below are the ones I know off by heart, and get recorded at time of Shut Down, so in theory, they could be sent up to Wargaming as a better indication as to why the player has suddenly dropped. 

Am I saying that this occurs, no, well it depends, I've never acutally looked at the code in WG games, nor do I care about it that much. 

BUT 🙂 Yes, windows does track states and why it was shut down. 

Additionally - 
 

Event ID Description
1074 Logged when an app (ex: Windows Update) causes the system to restart, or when a user initiates a restart or shutdown.
6006 Logged as a clean shutdown. It gives the message "The Event log service was stopped".
6008 Logged as a dirty shutdown. It gives the message "The previous system shutdown at time on date was unexpected".

 

In addition we have the additional logs that can determine action done by a user. 

4647 - Log off and Sign Out

For the life of me, i also can't remember the event ID number that tells you the Network State as well, not to mention battery health and sleep modes for Windows. But they are out there and if i worried more about it i would search. 

So as for your statement, Windows 100% knows if it was a clean power down / log off / application or a dirty shutdown. What It doesn't know is the 'exact' reason why and if you know the server family of windows, that's why when you boot back in, it will always prompt you for a 'why' and what happened. 

 

I think you were talking about the machine that was running the program in question. I do not think we were talking about the same thing though. The machine always knows why it was turned off, or abruptly shut down. I didn't say it can't cause its pretty obvious. Here we were talking about connectivity issues.

Let's say, you are machine A that was running WoWS. Your ISP is machine B. In between there are many NICs and NATs. Your system definitely knows what happened to your machine. It will never know what happened to machine B unless machine B decides to let you know, which sometimes it does before a restart. I am not certain about all the situations when that happens, so won't go there too far. Otherwise that is a definite breach of security. In the same way, lets say there is a router in between that connects to your pc via wire. And also connects to machine B with wires. You can go and pull the cable out, or a thunder strike could happen. or for some reason, machine B could simply shut down cause of crash or power failure. All you will see from machine A is no-connection to internet, or a limited access. Now you can connect to the router's external API's and take a look at what happened, I am pretty sure the game would not be able to do so, nor the OS will care much about it.

Now for the local machine (or your own machine), there is a thing about OS. OS knows many things that take place in karnel space, doesn't mean OS will want to pass that info over to user space where the programs are actually running, and for good reasons, security. Also, make no mistake, I can fool windows that it crashed pretty easily. I'm sure you can do that too. When you know how things happen, it's not that hard to replicate.

However, at the end of the day, if someone loses internet connection and cannot join in a game, its all on him. Its like you were appearing at an online exam and suddenly internet goes down. They won't stop the exam for you. Other cases, there are plenty of times to reconnect during the time-frame that would give anyone pink status.

Edited by icy_phoenix

Share this post


Link to post
Share on other sites
Member
100 posts
1,446 battles

you should be able see the internet route table of how your machine connects to wargaming Asia.

when then game started. There are 3 IP and ports connected. if you run netstat -tabn (run as admin)

netstat.thumb.jpg.9cd208c058b86e9887fdb20805156ac0.jpg

tracert those IPs,  and I can see the hops between my machine and WG Asia internet interface. There are enough information to  reflect the active segment and gateway in the whole data path or where it actually failed.

 

tracert.thumb.jpg.3a4dac0d30066d3ff109cd24c0335d15.jpg

They may not be all the machines in the data path. but it is enough to see how far my internet connection reach out to get to the destination.

Share this post


Link to post
Share on other sites
Super Tester
7,897 posts
8,143 battles
1 hour ago, tsuenwan said:

you should be able see the internet route table of how your machine connects to wargaming Asia.

when then game started. There are 3 IP and ports connected. if you run netstat -tabn (run as admin)

netstat.thumb.jpg.9cd208c058b86e9887fdb20805156ac0.jpg

tracert those IPs,  and I can see the hops between my machine and WG Asia internet interface. There are enough information to  reflect the active segment and gateway in the whole data path or where it actually failed.

 

tracert.thumb.jpg.3a4dac0d30066d3ff109cd24c0335d15.jpg

They may not be all the machines in the data path. but it is enough to see how far my internet connection reach out to get to the destination.

Firstly, thank's for the informative previous post. Now from this, how can you tell whose fault was it? And why do you think the game client connectivity always will follow a similar path instead of simply opening a tunnel that goes through some other routes? And some of these IPs are very well be NAT IP or virtual IPs.

It is true that game could have checked how many hops it could reach in the network, but what difference does it make that it falls right after your home router, or the local isp router, or the router next to it? They are all essentially the same output from system API. (you can test that as well).

So if I want to recap (lets see if we can agree on)

(a) Complete drop offs result in same regardless of accident or intention

(b) Link failures result in same regardless of accident or intention

(c) Router shut downs after leaving your home connection point should not stop you from reaching server unless you get a wide area blackout, cause there are usually redundant paths. But from this point till it reaches game servers, regardless of where it fails, they all are essentially same incident. Regardless of *where* OS simply sees a disconnect or no route, and then you can investigate further by traversing the route. OS has a set of states to describe network connectivity and it doesn't change based on where something happened as long as it is the same incident. This will be even more complicated when you consider the presence of double nats, cause they have their own issues in maintaining uninterrupted connections.

(d) Checking states in NIC... well that's what they can do, can you tell the NIC states that's somewhere along the route from your local machine? Sure you can use commands like "ip addr" to check network interface status in your system, how do you do that for a remote one? Some of them do have telnet based APIs but that requires authentication which you won't have. In other words, you can't, let alone a user space program running on your machine. Unless you decide to take a drive to the physical location of that machine and then access there.

So when you see no route in your system while playing games, then you traverse the route to see where it stopped and it usually ends up somewhere long before reaching destination, and it helps how? For example I have 2 hops even before leaving my home network (not uncommon for people using wifi subnets), but game doesn't know those hops are my property and I have control over them. So I can shut down the 2nd point or pull the plug there. Good luck telling the difference there. Anyway, how can you tell you were dropped off from server? or your OS for that matter. In OS you will see a terminated connection, and I am pretty sure a data center simply does not terminate a connection without a reason. It can be that the destination data center / server was shut down, but then all players were affected, which happened few times. And in that case server is aware and behaves accordingly. In the same way, it can be due to termination of any of the machines along the path. Or something (IP/port) changed during a packet transfer (which can happen) and server simply didn't recognize (which usually results in re-transmit but that can result in an expired time situation).

So even when you write a post disconnect investigation routine, it will almost always show you that something went wrong along the path before reaching destination, duh.

Now from a game engine's perspective, it can check network states from the OS api and see what state it is in. And then you want to run the tracing just to see that the connection breaks somewhere inbetween which it already knows? Which may very well be different from the time of incident?

This allowed me to come to conclusion, the game wont be able to recognize why it cannot connect to server apart from seeing that it cant. Surely it can find out where it failed, that is not helpful to begin with. And if the point of failure is the server itself, it already knows that and game does not need to find that for server. Wargaming may own the server, they don't own the connecting components in between and its on the user.

Edited by icy_phoenix

Share this post


Link to post
Share on other sites
Super Tester
7,897 posts
8,143 battles

TL;DR;

When OS sees something happened in connectivity, it simply reflects that by changing some of the states that you can see by using some APIs. Then you can do some investigating stuffs ofcourse. Now, the level of permission that requires on the system, I would not give that to WoWS client. I don't know why you would want to do that too.

Share this post


Link to post
Share on other sites
Member
100 posts
1,446 battles

 

3 hours ago, icy_phoenix said:

Firstly, thank's for the informative previous post. Now from this, how can you tell whose fault was it? And why do you think the game client connectivity always will follow a similar path instead of simply opening a tunnel that goes through some other routes? And some of these IPs are very well be NAT IP or virtual IPs.

It is true that game could have checked how many hops it could reach in the network, but what difference does it make that it falls right after your home router, or the local isp router, or the router next to it? They are all essentially the same output from system API. (you can test that as well).

So if I want to recap (lets see if we can agree on)

(a) Complete drop offs result in same regardless of accident or intention

(b) Link failures result in same regardless of accident or intention

(c) Router shut downs after leaving your home connection point should not stop you from reaching server unless you get a wide area blackout, cause there are usually redundant paths. But from this point till it reaches game servers, regardless of where it fails, they all are essentially same incident. Regardless of *where* OS simply sees a disconnect or no route, and then you can investigate further by traversing the route. OS has a set of states to describe network connectivity and it doesn't change based on where something happened as long as it is the same incident. This will be even more complicated when you consider the presence of double nats, cause they have their own issues in maintaining uninterrupted connections.

(d) Checking states in NIC... well that's what they can do, can you tell the NIC states that's somewhere along the route from your local machine? Sure you can use commands like "ip addr" to check network interface status in your system, how do you do that for a remote one? Some of them do have telnet based APIs but that requires authentication which you won't have. In other words, you can't, let alone a user space program running on your machine. Unless you decide to take a drive to the physical location of that machine and then access there.

So when you see no route in your system while playing games, then you traverse the route to see where it stopped and it usually ends up somewhere long before reaching destination, and it helps how? For example I have 2 hops even before leaving my home network (not uncommon for people using wifi subnets), but game doesn't know those hops are my property and I have control over them. So I can shut down the 2nd point or pull the plug there. Good luck telling the difference there. Anyway, how can you tell you were dropped off from server? or your OS for that matter. In OS you will see a terminated connection, and I am pretty sure a data center simply does not terminate a connection without a reason. It can be that the destination data center / server was shut down, but then all players were affected, which happened few times. And in that case server is aware and behaves accordingly. In the same way, it can be due to termination of any of the machines along the path. Or something (IP/port) changed during a packet transfer (which can happen) and server simply didn't recognize (which usually results in re-transmit but that can result in an expired time situation).

So even when you write a post disconnect investigation routine, it will almost always show you that something went wrong along the path before reaching destination, duh.

Now from a game engine's perspective, it can check network states from the OS api and see what state it is in. And then you want to run the tracing just to see that the connection breaks somewhere inbetween which it already knows? Which may very well be different from the time of incident?

This allowed me to come to conclusion, the game wont be able to recognize why it cannot connect to server apart from seeing that it cant. Surely it can find out where it failed, that is not helpful to begin with. And if the point of failure is the server itself, it already knows that and game does not need to find that for server. Wargaming may own the server, they don't own the connecting components in between and its on the user.

  1. The shown hops  are the internet segments ( which usually had their respective business owners) between my machine and the WG server. when data can not find its way from one segment to another, the routing report will show where/which segment it failed to hop to the next, and what gateways/exists it had tried but failed. in fact there already a few in my picture tough it found another way through. from that gateway's IP and you may find the operator of that IP from the internet registry. After all, I am not expecting WG to fix any internet problem for users, but just reflect the situation, and the  users shall be responsible in such cases for the lost of connection neither even they exit from the game.
  2. Then it come to when a user exist during a battle, or someone faking a crash,  then there are ways to track and distinguish between
  • Game crash, game hanging,
  • Game process being killed.
  • Global network performance degrades.
  • Local connection unplugged, Modem switch off.

These do not require complex computation nor consuming a lot of space, but reading and keeping some status data for a period of time.

3. When global connection performance degrades, there are ways to alert the server and the users to due with such situation more gracefully. and I am currently running multi-sites systems could operate in degraded, segmented and localized mode rather than just crash the user out when networking ran into trouble. and my first encounter of such a systems was operated in Australia for more than 30 years.

4. Please do not ignore that the network loading is the heaviest on the server-side, even the internet works perfectly. There could be mean to reflect their network performance and their process loading to warn user advance when they joining the game. at this moment, I can only read the udp drop rate for the interface and compare it with internet connection test result with other site from japan.

 

Edited by tsuenwan

Share this post


Link to post
Share on other sites
Super Tester
7,897 posts
8,143 battles
46 minutes ago, tsuenwan said:

 

  1. The shown hops  are the internet segments ( which usually had their respective business owners) between my machine and the WG server. when data can not find its way from one segment to another, the routing report will show where/which segment it failed to hop to the next, and what gateways/exists it had tried but failed. in fact there already a few in my picture tough it found another way through. from that gateway's IP and you may find the operator of that IP from the internet registry. After all, I am not expecting WG to fix any internet problem for users, but just reflect the situation, and the  users shall be responsible in such cases for the lost of connection neither even they exit from the game.
  2. Then it come to when a user exist during a battle, or someone faking a crash,  then there are ways to track and distinguish between
  • Game crash, game hanging,
  • Game process being killed.
  • Global network performance degrades.
  • Local connection unplugged, Modem switch off.

These do not require complex computation nor consuming a lot of space, but reading and keeping some status data for a period of time.

3. When global connection performance degrades, there are ways to alert the server and the users to due with such situation more gracefully. and I am currently running multi-sites systems could operate in degraded, segmented and localized mode rather than just crash the user out when networking ran into trouble. and my first encounter of such a systems was operated in Australia for more than 30 years.

4. Please do not ignore that the network loading is the heaviest on the server-side, even the internet works perfectly. There could be mean to reflect their network performance and their process loading to warn user advance when they joining the game. at this moment, I can only read the udp drop rate for the interface and compare it with internet connection test result with other site from japan.

 

Ok, fair enough, so this is what we can pick off

(1) As you said, WG can't fix these problems

(2) These local events are entirely on user and for the crash issue, unless its a problem faced by a good number of players, that won't raise a flag that something went wrong, instead, it will tell that something is not alright in client's machine.

(3) and (4) well I agree that WG could show server connectivity strength in the port in the form of a graph  so that players could see the pattern. Apart from that, unless they are being a cheapskate, their servers should be capable of handling the load pretty easily.

Share this post


Link to post
Share on other sites
Moderator
592 posts
2,995 battles
11 hours ago, icy_phoenix said:

Ok, fair enough, so this is what we can pick off

(1) As you said, WG can't fix these problems

(2) These local events are entirely on user and for the crash issue, unless its a problem faced by a good number of players, that won't raise a flag that something went wrong, instead, it will tell that something is not alright in client's machine.

(3) and (4) well I agree that WG could show server connectivity strength in the port in the form of a graph  so that players could see the pattern. Apart from that, unless they are being a cheapskate, their servers should be capable of handling the load pretty easily.

If my post seemed like I was directly poking you, that wasn't the intent. 

Between OS level information, Route Tables, Log Sends and a few other things, it's entirely possible to determine where in the stack it's failed, but it's complex and not something I would expect a game to be able to do. 

So yes....

1. WG cannot fix internet issues, that's down to the mega corps who own the internet routes

2. Local crashes, could, in theory be sent up to the game to say if it was normal / abnormal, but I would Totally not want that to occur, sets a precedent that I wouldn't want to see. 

3. I think you'll find the servers are more than capable of handling the load, the issue will be more down to the bandwidth available at a number of points 😉

Share this post


Link to post
Share on other sites
Member
100 posts
1,446 battles
6 hours ago, tc1259 said:

If my post seemed like I was directly poking you, that wasn't the intent. 

Between OS level information, Route Tables, Log Sends and a few other things, it's entirely possible to determine where in the stack it's failed, but it's complex and not something I would expect a game to be able to do. 

So yes....

1. WG cannot fix internet issues, that's down to the mega corps who own the internet routes

2. Local crashes, could, in theory be sent up to the game to say if it was normal / abnormal, but I would Totally not want that to occur, sets a precedent that I wouldn't want to see. 

3. I think you'll find the servers are more than capable of handling the load, the issue will be more down to the bandwidth available at a number of points 😉

Network diagnostic skills may not be essential for normal programmers who develop simple business applications. but it is one of the core skill for game developers/designers making MMORPG Online games.

It is not that complex for a game engine to perform basic network diagnostic. During early days of internet, In order to maintain a group game, It was even more essential to facilitate sufficient critical game data delivery via unreliable network/medium at the time. Games like Red Alert, AOE, or Diablo II  had game design to adapt to very different data round trip performance, Network condition monitoring and handling was one of the essential function then and is still now.  and books like "Foundations of Online Games" by Kengo Nakajima had dedicated 40% of the content explaining how a game could work with various network conditions and reliability levels.

then:

1. It is true WG can not fix internet problem. but user crashed out of battle due to network issues. then there are evidences to inform or explain to his/her team members about the situation.

2. Just like windows or other software dealing with crash log during their recovery. users could have the control to sent or not sent. but it helps to find out who is responsible.

3. It will be more convincing if the users are informed of the server and its immediate connection situation in the game. Hardware CPU and OS may be very reliable to handle loads. How software handles loads are usually another problem.

 

Edited by tsuenwan

Share this post


Link to post
Share on other sites
Sign in to follow this  

×