Possible fix for lag comp/bufferbloat

[Netduma Iain]

Hi guys, I have been doing some testing on black ops 3 using a custom Linux router (openwrt), by tweaking the queues of all router interfaces using pfifo queuing managment, limiting each queue to 3 packet length drastically improved my online experience.

 

I'm finding that most games im finishing top of the leaderboard, with a kd of 2.89 at present.  If something like this can be implemented with netduma i'm sure the results would be even better.

 

The only conclusion i can come to is that the majority of problems in online FPS games is in fact caused by bufferbloat, throttling my upload made little to no difference.

 

Hi Roy,

 

It's very cool to see you learning about OpenWRT/Linux and networking.

 

However I must add that buffer-bloat is a very very complicated subject. We have solved it near optimally on our router. Just use anti-flood 70% each way with reactive/pre-emptive.  I say near optimally, because we have the best solution (parent pending) that we will release soon to the public. It will be pre-emptive algorithm on steroids

 

As I said it's a very complicated problem so naturally people will make mistakes if they haven't spent thousands of hours researching it. I'll address a few of the incorrect assertions you made:

• Bufferbloat is unlikely making any difference at all if no other devices are downloading/uploading at your house. So in a normal game bufferbloat won't be making much difference.
• K/D is a very bad metric for measuring network performance it's a far better metric for sports psychology. If you do a fair test you'll probably find no correlation between K/D and buffer bloat
• Finally settings packet length to 3 is a terrible idea it will make your game much worse as it's far more likely you'll drop gaming packets. Setting a short queue length is intuitively a good idea but it's actually very counter-productive not only on game traffic but all traffic. All the research going in to buffer-bloat and queue theory is basically about deciding when a "queue" should drop a packet and setting it small has been proven to be a very bad choice
• PFifo is what hyper-lane is essentially doing, so you already have that technology available to you
• Finally setting up qdisc is very tricky process for a leymen and I've seen it hundreds of times where it's misconfigured and it's not actually doing anything at all. Make sure you test your settings using ping and varying traffic profiles.

Our results from the lag comp experiment will be coming out soon, make sure to stick about as I'm sure they'll be of interest to you

 

Happy tinkering! 

[slingblade5]

Hi Roy,

 

It's very cool to see you learning about OpenWRT/Linux and networking.

 

However I must add that buffer-bloat is a very very complicated subject. We have solved it near optimally on our router. Just use anti-flood 70% each way with reactive/pre-emptive.  I say near optimally, because we have the best solution (patent pending) that we will release soon to the public. It will be pre-emptive algorithm on steroids

 

As I said it's a very complicated problem so naturally people will make mistakes if they haven't spent thousands of hours researching it. I'll address a few of the incorrect assertions you made:

• Bufferbloat is unlikely making any difference at all if no other devices are downloading/uploading at your house. So in a normal game bufferbloat won't be making much difference.
• K/D is a very bad metric for measuring network performance it's a far better metric for sports psychology. If you do a fair test you'll probably find no correlation between K/D and buffer bloat
• Finally settings packet length to 3 is a terrible idea it will make your game much worse as it's far more likely you'll drop gaming packets. Setting a short queue length is intuitively a good idea but it's actually very counter-productive not only on game traffic but all traffic. All the research going in to buffer-bloat and queue theory is basically about deciding when a "queue" should drop a packet and setting it small has been proven to be a very bad choice
• PFifo is what hyper-lane is essentially doing, so you already have that technology available to you
• Finally setting up qdisc is very tricky process for a leymen and I've seen it hundreds of times where it's misconfigured and it's not actually doing anything at all. Make sure you test your settings using ping and varying traffic profiles.

Our results from the lag comp experiment will be coming out soon, make sure to stick about as I'm sure they'll be of interest to you

 

Happy tinkering! 

 

 

 

 

 (I just pee'd a lil bit sitting here reading that)       Awesome stuff, very excited for the future with the Netduma as always.

[Netduma Iain]

heh coming realtively soon

[richb-hanover]

Hi Roy,

 

It's very cool to see you learning about OpenWRT/Linux and networking.

 

However I must add that buffer-bloat is a very very complicated subject. We have solved it near optimally on our router. Just use anti-flood 70% each way with reactive/pre-emptive.  I say near optimally, because we have the best solution (parent pending) that we will release soon to the public. It will be pre-emptive algorithm on steroids

 

As I said it's a very complicated problem so naturally people will make mistakes if they haven't spent thousands of hours researching it. I'll address a few of the incorrect assertions you made:

• Bufferbloat is unlikely making any difference at all if no other devices are downloading/uploading at your house. So in a normal game bufferbloat won't be making much difference.
• K/D is a very bad metric for measuring network performance it's a far better metric for sports psychology. If you do a fair test you'll probably find no correlation between K/D and buffer bloat
• Finally settings packet length to 3 is a terrible idea it will make your game much worse as it's far more likely you'll drop gaming packets. Setting a short queue length is intuitively a good idea but it's actually very counter-productive not only on game traffic but all traffic. All the research going in to buffer-bloat and queue theory is basically about deciding when a "queue" should drop a packet and setting it small has been proven to be a very bad choice
• PFifo is what hyper-lane is essentially doing, so you already have that technology available to you
• Finally setting up qdisc is very tricky process for a leymen and I've seen it hundreds of times where it's misconfigured and it's not actually doing anything at all. Make sure you test your settings using ping and varying traffic profiles.

Our results from the lag comp experiment will be coming out soon, make sure to stick about as I'm sure they'll be of interest to you

 

Happy tinkering! 

 

Hi Iain, It's good to see another router vendor who's addressing the problem of bufferbloat. (So few vendors even seem to understand the problems it causes...) I've been working with the CeroWrt team to get rid of it. Fortunately, it has been available for three years in OpenWrt. Some comments on your observations:

• Yes, bufferbloat is complicated. That's why we also spent thousands of hours researching ways to make an effective algorithm (fq_codel) very simple to use, so people don't screw it up. (Basically, it has two parameters - upload and download speed).
• I will disagree that bufferbloat isn't a factor when no one else is up/downloading. Today's web pages contain a lot of data - averaging 2 mbytes per page. (Not that long ago, a two megabyte file would have been considered a big transfer :-) Even today, it's a significant amount of data, and if your router is allowing your packets (gaming, VoIP) to be queued up behind someone else's web traffic, you're going to see lag. 
• I will agree with you that a hard limit on the number of packets in a queue is a bad strategy. It's far better to let the router's smart queue management determine how long packets have dwelled in the queue, and offer flow control (or drop packets) for flows that are using more than their share of the bandwidth.
• I'm glad you mentioned testing. As a quick test, I always recommend www.dslreports.com/speedtest. It measures down/upload speeds, but also measures the latency/lag *during* the transfers. This is where you will frequently see bufferbloat. 
• But for a better test, the bufferbloat team has created an accurate and repeatable test platform for testing out the various algorithms that we devised. It's called Flent (FLExible Network Tester), and you can get it at www.flent.org. When you are ready to talk about your algorithm, I hope you will publish the results of a Flent test, so it will be possible to compare with the established bufferbloat control algorithms. 
• I assume you meant to say that your pre-emptive algorithm is "patent pending". Congratulations. It will be all the more interesting to see how it compares to fq_codel. 

 

Best regards,

Rich Brown

[Netduma Iain]

Hi Iain, It's good to see another router vendor who's addressing the problem of bufferbloat. (So few vendors even seem to understand the problems it causes...) I've been working with the CeroWrt team to get rid of it. Fortunately, it has been available for three years in OpenWrt. Some comments on your observations:

• Yes, bufferbloat is complicated. That's why we also spent thousands of hours researching ways to make an effective algorithm (fq_codel) very simple to use, so people don't screw it up. (Basically, it has two parameters - upload and download speed).
• I will disagree that bufferbloat isn't a factor when no one else is up/downloading. Today's web pages contain a lot of data - averaging 2 mbytes per page. (Not that long ago, a two megabyte file would have been considered a big transfer :-) Even today, it's a significant amount of data, and if your router is allowing your packets (gaming, VoIP) to be queued up behind someone else's web traffic, you're going to see lag. 
• I will agree with you that a hard limit on the number of packets in a queue is a bad strategy. It's far better to let the router's smart queue management determine how long packets have dwelled in the queue, and offer flow control (or drop packets) for flows that are using more than their share of the bandwidth.
• I'm glad you mentioned testing. As a quick test, I always recommend www.dslreports.com/speedtest. It measures down/upload speeds, but also measures the latency/lag *during* the transfers. This is where you will frequently see bufferbloat. 
• But for a better test, the bufferbloat team has created an accurate and repeatable test platform for testing out the various algorithms that we devised. It's called Flent (FLExible Network Tester), and you can get it at www.flent.org. When you are ready to talk about your algorithm, I hope you will publish the results of a Flent test, so it will be possible to compare with the established bufferbloat control algorithms. 
• I assume you meant to say that your pre-emptive algorithm is "patent pending". Congratulations. It will be all the more interesting to see how it compares to fq_codel. 

 

Best regards,

Rich Brown

 

Hi Rich,

 

Have we chatted before here? I tried to follow up a dialogue with someone from CeroWRT a while a go but I think they didn't return to the forum. Would be great to see if we can work together. Right just quickly to clarify my responses:

• I understand the codel is a simple solution hence it's motto no knobs, what I meant is the subject matter is complicated. So trying to come up with your "own solution" will likely be perilous unless you fully understand networking buffer bloat and related queue and control theory. 
• When speaking here I assume most of our customers are not experts in networking. So I was trying to make it clear that unless something else is using the Internet buffer bloat is not an issue. We know full well about micro-spikes - that's what our optimised algorithm solves brilliantly. The interesting thing about micro-spikes is that most games use a moving average to measure ping so you in general will never see them visible on a games ping graph/number
• Agreed, the whole field of active-queue-management and more specifically bufferbloat aimed algorithms is based on this. It would be pretty embarrassing if the solution was simply reduce the queue size
• On testing, we recommend customers just set their anti-flood to 70% and that should solve 99% of problems. This is a support forum after all
• Finally I did mean patent pending. Tbh we'd prefer not to go that way but we're a small company and need to protect ourselves some way. As we get larger we will probably consider open sourcing it.

 

If you like let's exchange emails, we could get you to test our new algo when we release it

[blakeyb1]

If we don't know how cods lag comp code works what's to say lowering the que size isn't the fix though, I for one am using my BT homehub again cause I am getting absolutely bsed in p2p matches against Germans and the like.

[A7Legit]

If we don't know how cods lag comp code works what's to say lowering the que size isn't the fix though, I for one am using my BT homehub again cause I am getting absolutely bsed in p2p matches against Germans and the like.

Netduma had a lag comp test with BO3 a few days ago, they tested this and will reveal results when they're ready.

[RoyDavis77]

Thanks for the info Iain, I get what you're saying, but this just seems to work with black ops 3, don't know why, 5 days now of consistent slaying and lobby leading, I have noticed with tweaking the queuing that my ping on both wired and wireless has dropped, im getting 19ms ping to Germany on dslreports, before it would be between 38 and 45ms, and that was on a wireless connection, also averaging 10 to 15ms of buffering on the download and upload, before it would go as high as 3000ms.

 

Cant wait to see the results of your lag comp test, will be eye opening for sure.

[blakeyb1]

Netduma had a lag comp test with BO3 a few days ago, they tested this and will reveal results when they're ready.

I'm not trying to say their way of thinking is wrong because it isn't a stable connection should = good bullet reg but on this game it just isn't like that.

 

Roy can you dm me what I'd need to do/buy to do this setup you're doing since I'm not all that router savvy but can easily learn but don't wanna really brick my router tbh .

 

Edit: I can't wait for the results to, since I was a part of the test myself .

[blakeyb1]

Can anybody who's proficienct with openwrt help me set these settings up?

[mr.victor]

please share your settings as i just got openwrt and am new to this. thank you

[kritikal]

Can anybody who's proficienct with openwrt help me set these settings up?

 

If you have openwrt installed all you need to do is SSH in with a program like putty and enter the commands

[Matt_Akira]

Download putty from Google when you open putty type root then your password it will ssh into openwrt then type your commands

[Sleightofdean]

So I've just downloaded Network Emulator Client for Windows and running connection through my PC to PS4.

 

You can add artificial latency, alter loss, error, queuing etc i'm guessing this is kind of the same thing as what you are doing with openwrt but a more basic Windows version?

 

Not played around with it too much yet as i only got it last night and my knowledge of networking isn't great so still trying to figure out how some of it works.

[Netduma Fraser]

We don't add any artificial latency or anything that would be considered cheating.

[Torkild]

The lag comp test... haven't that been  "very soon released" for about a couple of  months  already?  
 
Any whore, it will be interesting to see the result! And even better if they finally made a conclusion and some thing

they can apply in firmware so that the netduma makes a BIG impact in the way the game is played

 As the game (BO3) is total randomness at this point..  On normal weekdays connection and hit registration is spot on for about 50% of the time. 
30% its sketchy .. you now.. It takes a couple of shots before the first register on the other player.. but it seems to affect only  1-2 player on the opposing team.. 
10% It's so messed up that you shot first, get one hit registered and your then instant killed. That's latency F*cing you over so hard that the other player saw you first and probably shot 3 times before you got the first hit-marker on him.. And when his final bullet hit, it feels like you getting all the damage at once..  And then there's 10% where the entire game just have random hick ups, the game stops when moving around, rubber banding and if you check the ping bar /score menu.. there's just a disco  

On weekends it's even worse.  30% Perfect, 50% sketchy, 20% Rubber-banding ping disco party.  
 
 

[srmiller82]

 

To the OP, a lot of the info is ridiculous, first off Black OP 3 and every other COD doesn't not have dedicated servers if you heard this they lied, as they refuse to pay for this as it is quite expensive, the only shooter as of today to have those is halo 5 no other shooter has it, 2nd the reason your k/d is so good when you changed those QOS setting yours connection is lagging so bad to everybody else that it forces BLOPS3 to give you host, as this is a p2P game, not only do you have host but its next to impossible for anybody else to kill you its borderline cheating, this type of behavior in a game like destiny will get you banned, but in essence your forcing host and then lagging it so bad that no bullets ever hit you, you are not doing better your basically cheating and ruining the game for everyone else in the lobby, case and point when you said you were in a bad lobby where everybody was yellow and red bar, yes that is their connection to YOUR lobby as you are host and it lags that bad for them not because of their connection but because of yours, taking off QOS will make you very very easy to kill as you should be, this isn't anything against QOS just the way you have it set making the game awful for everyone in the lobby. And 3rd, you said you have a 80MBPS fiber connection, QOS is completely unnecessary, because 100 people couldn't saturate that line enough to cause lag in black ops 3, you do not need QOS, I thought read someone in here saying that web pages can be content intensive yes, they can but unless you have 200 pages open it wont saturate 80Mbps, and I'm very very very versed in the topic of p2p and servers and FPS with QOS, I haven't spent hours optimizing my connection to make it not lag online, I also have fiber but its only 10Mbps so it saturates 8x quicker than yours, what I have found is when I'm not lagging in the slightlest and being on fiber makes me extremely easy to kill because all bullets land on p2P connections, and makes you feel like you suck, that's because 80% peoples connections lag and when you have a very good one you are at a disadvantage, but doesn't give you the right to force host and make everybody lag so you can get your 3.0/kd that isn't deserved.

[K4L4M1TY]

 

 

To the OP, a lot of the info is ridiculous, first off Black OP 3 and every other COD doesn't not have dedicated servers if you heard this they lied, as they refuse to pay for this as it is quite expensive, the only shooter as of today to have those is halo 5 no other shooter has it, 2nd the reason your k/d is so good when you changed those QOS setting yours connection is lagging so bad to everybody else that it forces BLOPS3 to give you host, as this is a p2P game, not only do you have host but its next to impossible for anybody else to kill you its borderline cheating, this type of behavior in a game like destiny will get you banned, but in essence your forcing host and then lagging it so bad that no bullets ever hit you, you are not doing better your basically cheating and ruining the game for everyone else in the lobby, case and point when you said you were in a bad lobby where everybody was yellow and red bar, yes that is their connection to YOUR lobby as you are host and it lags that bad for them not because of their connection but because of yours, taking off QOS will make you very very easy to kill as you should be, this isn't anything against QOS just the way you have it set making the game awful for everyone in the lobby. And 3rd, you said you have a 80MBPS fiber connection, QOS is completely unnecessary, because 100 people couldn't saturate that line enough to cause lag in black ops 3, you do not need QOS, I thought read someone in here saying that web pages can be content intensive yes, they can but unless you have 200 pages open it wont saturate 80Mbps, and I'm very very very versed in the topic of p2p and servers and FPS with QOS, I haven't spent hours optimizing my connection to make it not lag online, I also have fiber but its only 10Mbps so it saturates 8x quicker than yours, what I have found is when I'm not lagging in the slightlest and being on fiber makes me extremely easy to kill because all bullets land on p2P connections, and makes you feel like you suck, that's because 80% peoples connections lag and when you have a very good one you are at a disadvantage, but doesn't give you the right to force host and make everybody lag so you can get your 3.0/kd that isn't deserved.

 

That right there is the problem man, this is my issue, I firmly believe the better your upload and ping/jitter are, the easier you get melted and that's completely unfair to begin with. I play well and i'm good but when I die from someones weaker equipped gun in a half second burst whilst they take four hitmarkers out of a whole clip in which I know almost every bullet hit i'm calling them out on it. I'm sure everyone else notices the 3 bars always at the top of the scoreboards.

 

Back to what somebody said in an earlier post(think it was a Duma employee/rep) added jitter being an issue. One major thing I notice in lobbies (i always connect to 'e servers') is when everyone is on a solid four bar connection they all die the same and very quickly but as soon as a 2-3-glitchy flickering 3-4 bar joins the lobby turns to sponge heaven, the easy kills still die from being outskilled but these magical connections that join seem to take way more hit markers or not register very well. I come to the conclusion if there's any signs of a player not having a solid 4 bar connection in the lobby it's bound to have some sketchiness.

 

One last thing today the game has been messed up worse than i've seen yet, on at least two separate occassions I was boost sliding then all of a sudden i'm boost sliding the opposite direction, these servers are becoming a nightmare.

[bagsta69]

OK, so I'm not sure what exactly is getting said here. Black Ops and several Call of duty games before it have had dedicated servers. I think most people know this is a given.

Everybody's  experience will differ depending on their own circumstances but in my own personal case I have experienced really bad games for a few weeks with high jitter and packet loss due to my ISP line having issues at the white box and vdsl plate connected to it.

As soon as this was resolved and I had near to zero jitter and near zero packet loss it was like playing a different game. 70% of my games over a 2-3 hour gaming session I was in the winners circle with my congestion control set at 70/70. which gives me about 16mb down and 4mb up. 

Double XP weekends still throw up a few issues but that would also point to there being dedicated servers as they get over congested.

When you switch your game on and look at the Geo map you can see the location of these dedicated servers.

As I said earlier though, everybody's experience will differ but if you are having issues it may worth running a long pingplotter test to check your line isn't at fault.

I swore mine was fine until I did this and discovered it was not. Difficult to tell as well as all other internet related activities worked fine.

Cheers yall, Happy Hunting.

[Net-1]

I would like to second the importance of sorting out your jitter and packet loss

 

I have a high bandwidth low ping connection but with out netduma it runs high jitter with some packet loss.

 

With duma at cc 70/70 and Hyperlane active my jitter and packet loss vanish and games are crisp.

 

I think when your modem is pushing your allowed bandwidth to the max, which it will always try to do, it introduces packet loss and jitter. For most users this is a non issue as they are mostly concerned with bandwidth. But for real time network use like gaming you need to minimise this.

 

Cod developers have stated repeatedly that lobbies are based on connection quality first over everything else.

 

It is very likely that if you are continually being put into laggy games it is because your connection is not as fast or stable as you think and you have been connection matched with other high ping high jitter etc gamers.

[UK_Wildcats_Fans]

Cod developers have stated repeatedly that lobbies are based on connection quality first over everything else.

 

It is very likely that if you are continually being put into laggy games it is because your connection is not as fast or stable as you think and you have been connection matched with other high ping high jitter etc gamers.

 

They say it is based upon connection quality first, but there is an additional factor that may NOT be the best connection.  They try to match players with the same map packs FIRST and then try to match the best connection quality second.   As more DLC come out, there are a larger number of varying combinations of maps and lower player count.  For example, when BO3 first launched, every player was in the same match making pool base.  When DLC1 was released, you then have player base 1 without any DLC and player base 2 with DLC 1.   With DLC 2 we now have more player bases and less players per base.  For example, you have people without DLC in player base 1, DLC 1 people in player base 2, people with DLC 1 and DLC 2 in player base 3.     

[JConnor]

 

 

To the OP, a lot of the info is ridiculous, first off Black OP 3 and every other COD doesn't not have dedicated servers if you heard this they lied, as they refuse to pay for this as it is quite expensive, the only shooter as of today to have those is halo 5 no other shooter has it, 2nd the reason your k/d is so good when you changed those QOS setting yours connection is lagging so bad to everybody else that it forces BLOPS3 to give you host, as this is a p2P game, not only do you have host but its next to impossible for anybody else to kill you its borderline cheating, this type of behavior in a game like destiny will get you banned, but in essence your forcing host and then lagging it so bad that no bullets ever hit you, you are not doing better your basically cheating and ruining the game for everyone else in the lobby, case and point when you said you were in a bad lobby where everybody was yellow and red bar, yes that is their connection to YOUR lobby as you are host and it lags that bad for them not because of their connection but because of yours, taking off QOS will make you very very easy to kill as you should be, this isn't anything against QOS just the way you have it set making the game awful for everyone in the lobby. And 3rd, you said you have a 80MBPS fiber connection, QOS is completely unnecessary, because 100 people couldn't saturate that line enough to cause lag in black ops 3, you do not need QOS, I thought read someone in here saying that web pages can be content intensive yes, they can but unless you have 200 pages open it wont saturate 80Mbps, and I'm very very very versed in the topic of p2p and servers and FPS with QOS, I haven't spent hours optimizing my connection to make it not lag online, I also have fiber but its only 10Mbps so it saturates 8x quicker than yours, what I have found is when I'm not lagging in the slightlest and being on fiber makes me extremely easy to kill because all bullets land on p2P connections, and makes you feel like you suck, that's because 80% peoples connections lag and when you have a very good one you are at a disadvantage, but doesn't give you the right to force host and make everybody lag so you can get your 3.0/kd that isn't deserved.

 

This sounds like something the idiots on the official CoD forums would spew...

[AsNCo]

To the OP, a lot of the info is ridiculous, first off Black OP 3 and every other COD doesn't not have dedicated servers if you heard this they lied, as they refuse to pay for this as it is quite expensive, the only shooter as of today to have those is halo 5 no other shooter has it, 2nd the reason your k/d is so good when you changed those QOS setting yours connection is lagging so bad to everybody else that it forces BLOPS3 to give you host, as this is a p2P game, not only do you have host but its next to impossible for anybody else to kill you its borderline cheating, this type of behavior in a game like destiny will get you banned, but in essence your forcing host and then lagging it so bad that no bullets ever hit you, you are not doing better your basically cheating and ruining the game for everyone else in the lobby, case and point when you said you were in a bad lobby where everybody was yellow and red bar, yes that is their connection to YOUR lobby as you are host and it lags that bad for them not because of their connection but because of yours, taking off QOS will make you very very easy to kill as you should be, this isn't anything against QOS just the way you have it set making the game awful for everyone in the lobby. And 3rd, you said you have a 80MBPS fiber connection, QOS is completely unnecessary, because 100 people couldn't saturate that line enough to cause lag in black ops 3, you do not need QOS, I thought read someone in here saying that web pages can be content intensive yes, they can but unless you have 200 pages open it wont saturate 80Mbps, and I'm very very very versed in the topic of p2p and servers and FPS with QOS, I haven't spent hours optimizing my connection to make it not lag online, I also have fiber but its only 10Mbps so it saturates 8x quicker than yours, what I have found is when I'm not lagging in the slightlest and being on fiber makes me extremely easy to kill because all bullets land on p2P connections, and makes you feel like you suck, that's because 80% peoples connections lag and when you have a very good one you are at a disadvantage, but doesn't give you the right to force host and make everybody lag so you can get your 3.0/kd that isn't deserved.

 

The amount of assumptions and nonsense out of you and with this being the first post clearly says that you are more of a troll who doesn't even have a Netduma to even see the results for yourself.

If you do not study IT and more specifically on Networking, I would suggest you shut your trap instead of making so much noise like an empty bin that you are.

 

 

 

first off Black OP 3 and every other COD doesn't not have dedicated servers if you heard this they lied, as they refuse to pay for this as it is quite expensive, the only shooter as of today to have those is halo 5 no other shooter has it,

To clarify on this question, the Netduma team checks on the IP address that members sent to confirm that it is a dedicated server before implement it into the cloud update which will then appear on the Host Filtering page.

It took me a while but I finally figured out the IP addresses that I connect to when it comes to playing in the actual matches. And for Black Ops 3 specifically in my region, I will connect to GameServers.com (Japan). That is the dedicated game server that Activision is renting from for my region.

It may even come as a surprise for you but Black Ops 2 on the PS3 uses Dedicated Servers too but depending on how bad your connection is or whether you are using the Netduma to block the servers off, it is quite easy to get into a P2P connection. The same cannot be said for Black Ops 3 on the PS4 which will always try to put you in a dedicated server as much as possible unless you really have a terrible connection in which you'll be playing with other people with very bad connection and the experience is awful when I tried that method myself.

 

The reason why you may feel that the experience is almost the equivalent of a P2P connection is because Activision buys the cheapest servers and many PC players who have played on GameServers.com on their own games can testify on its poor performance which result in some people still experiencing lag on these servers. So you can thank Activision on their stinginess.

 

 

 

what I have found is when I'm not lagging in the slightlest and being on fiber makes me extremely easy to kill because all bullets land on p2P connections, and makes you feel like you suck, that's because 80% peoples connections lag and when you have a very good one you are at a disadvantage, but doesn't give you the right to force host and make everybody lag so you can get your 3.0/kd that isn't deserved.

The original purpose of the Netduma Host Filtering was specifically for P2P. All thanks to these laggy players, your own experience is ruined. Netduma solves this by giving you the chance to let you play with people within a certain distance or you can choose to play with people based on ping.

But it does not specifically give you advantage. It is actually the opposite. Now everyone in that lobby are close to one another. And that means you will all have lower ping so the game will be enjoyable for everyone. But too bad. By the time the Netduma is released, COD has started to move to Dedicated Servers so that part of the feature is only useful in getting you into the servers that you want but you can't really control who can get into that lobby. And that means, you might still have a terrible experience thanks to those laggers.

 

There are only 2 solutions to this. Rent more servers or go for really really high quality servers which can do more calculations to balance out the lag compensation.

 

Many people are looking forward to COD4 Modern Warfare remastered. But sadly, thanks to the confirmation that they are using Dedicated Servers again, the experience may not be the same as the P2P connection that you once had and might have enjoyed in the past but instead expect yourself to get a similar experience to Black Ops 3 for better or worst.

[iAmMoDBoX]

Many people are looking forward to COD4 Modern Warfare remastered. But sadly, thanks to the confirmation that they are using Dedicated Servers again, the experience may not be the same as the P2P connection that you once had and might have enjoyed in the past but instead expect yourself to get a similar experience to Black Ops 3 for better or worst.

 

Yup, that's why I'm not getting my hopes up... They always find a way to mess something good up.

[JConnor]

Also, MW was released in 2007.

 

Nine years later, and there are exponentially more people online across all mediums e.g. wi-fi, cable broadband, DSL, fiber, etc. So, this will not be like the original MW because there will be just too many people compared to the relatively small XBL and PSN networks of old...