• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Teaming

Both the NICs and the Switch have to support it, but the more important question is why. Unless you are saturating an already existing connection, why bother? Just because you get it to the switch twice as fast, the receiving system would also need to be teamed into the switch to get it twice as fast.
 
Gillbot said:
Both the NICs and the Switch have to support it, but the more important question is why. Unless you are saturating an already existing connection, why bother? Just because you get it to the switch twice as fast, the receiving system would also need to be teamed into the switch to get it twice as fast.
Click to expand...

Not to mention a single conversation will stay on the same link.

Teaming is useful for servers with lots of conversations to distribute the load. For one to one conversations it's worthless.
 
I wouldn't imagine most people would feel network congestion and how it effects online action in UT3. I've been playing it almost everyday the past 4-5yrs. I can feel a lag in my computer system due to something. I'm not 100% positive it's networking related but I have a good hunch. My RT N66U has a 600MHz processor onboard. Turning on QoS traffic manager disables it's hardware acceleration processor on the Router switch. While it does prioritize a list of packet types like ACK, SYN, ICMP and a few others there is a noticeable lag in play, and my computer feels like it's input response increases 10 fold.
 
How would teaming help with your lag? Teaming can solve bandwidth issue it can't solve a latency issue. If anything it'll created slightly higher latency.
 
Wait, are you saying that I'm saying it causes lag? No. If I did I didn't mean to cause it to sound that way. If by lag you also are using the term latency which I assume you mean ping, too, then no. Either someone is using a hack or there is a definite issue going on somewhere. I understand and can feel the difference between every "1" ACK and every "2nd" ACK. 2 will release CPU cycles from networking duties making input latency e.g. double tapping to dodge, more responsive like immediately. When there is more strain on the CPU when ACK is set to 1 the calls to say double tapping again, are not as responsive. More delayed. This is the type of thing that happens when Hardware Acceleration is disabled on the Router.
 
Your bottleneck is not on your gigabit lan, it's on your internet connection, unless you have a connection that's higher than gigabit... though bandwidth is not the only factor, latancy is, but your home network will have much less latancy than the internet, too.

Basically, nothing you do on your own network will improve your internet performance.
 
Teaming on regular End-User Network is just like:

"Help me to find a Harley engine that can be installed on my Vespa Scooter".

Can Not be done, and even if somehow can,it will not do anything for the Vespa.

piaggio-vespa-060112.jpg



😎
 
Shaving off a couple milliseconds of latency is not going to make you a better player. I don't care how sensitive you think you are.

Also, link aggregation protocols can actually ADD latency to your connection. They have to hash each and every packet that runs through them.

Also, Server 2012 can perform LACP independent of NIC.
 
BTRY B 529th FA BN said:
Wait, are you saying that I'm saying it causes lag? No. If I did I didn't mean to cause it to sound that way. If by lag you also are using the term latency which I assume you mean ping, too, then no. Either someone is using a hack or there is a definite issue going on somewhere. I understand and can feel the difference between every "1" ACK and every "2nd" ACK. 2 will release CPU cycles from networking duties making input latency e.g. double tapping to dodge, more responsive like immediately. When there is more strain on the CPU when ACK is set to 1 the calls to say double tapping again, are not as responsive. More delayed. This is the type of thing that happens when Hardware Acceleration is disabled on the Router.
Click to expand...

I'm not sure who fed you that Kool-aid, but those ACKs and router CPU cycles are not something you can "feel." They are happening considerably faster than you think they are. That's like saying you can see the pulses of light in a fiber optic cable with your naked eye. Nor is the TCP/IP stack somehow taxing your processor to the point where there's a visible slowdown, it's not capable of doing so.

Two Teamed Killer NICs aren't going to do much for you if the server you're connecting to is the problem. As others have said, there's not much you can do to manipulate the traffic coming in and out of your gateway device, and the NT-66 is probably the best and fastest SOHO router on the market. You're still at the mercy of your ISP and every router between you and that game server.

UT3 is not going to saturate even the crappiest home router's abilities, Gigabit Ethernet, or even the cheaper broadband connections unless you're literally trying to run dozens of UT3 servers out of your home.
 
BTRY B 529th FA BN said:
So let me get this straight, who is thinking I am talking about my ping to the server, and how I can lower it, raise your hand.
Click to expand...

You are either talking about your PING (whether you think you are or not) or you are talking about an issue with game lag on your PC (either a software or hardware issue on the computer). NIC teaming won't help either of those situations, and will actually make the first one slightly worse.
 
You said that you can 'feel' latency in the game through your connection to your router. Unless there is something physically wrong with your router/cables/NIC, or you are pushing so much bandwidth through the router that you are capping the limits of your Internet connection or overloading the router itself, that's impossible.

It IS possible that you can 'feel' latency on the connection from your computer all the way to the game server, but you can't do anything about that outside of your own router.
 
Last edited:
Gillbot said:
Unlikely

Probably
Click to expand...

Ding ding ding! We have a winner!

Teaming nics is something i do only on occasion due to most applications just not needing it, a gig link is hard to saturate on a LAN to the point where adding a second is any benefit. teaming for an internet connection is absolutely worthless unless you happen to have a multi-gig connection.

I personally love UT and play it all the time with no perceptible lag, and I am very picky when it comes to lag. Its either your connection or the server you are connected to.
 
This is a Trace from my computer to Epic UT server.

The only thing that is under my control is from the computer to the Router and that (step 1) shows response of 0 ms

From there to the server there is a lots of Ups and Downs.

However, none of them is under the End-User control.

1, 192.168.1.254, 0ms,MDS,0%

2, 72.225.224.1, 31ms,cpe-72-225-224-1.ny,0%

3, 24.29.139.62, 23ms,gig-0-3-0-12-nycmny,0%

4, 24.29.119.130, 13ms,tenge-0-4-0-7-300-n,0%

5, 184.152.112.107, 26ms,bun101.nycmnytg-rtr,0%

6, 24.29.148.250, 18ms,bun6-nycmnytg-rtr00,0%

7, 107.14.19.24, 16ms,None,0%

8, 107.14.19.153 ,21ms,None,0%

9, 69.31.94.57, 47ms,xe-5-0-6.ar2.ewr1.u,0%

10, 69.31.94.117, 30ms,ae3-80g.cr1.ewr1.us,0%

11, 69.22.142.74, 37ms,xe-5-0-0.cr1.ord1.u,0%

12, 69.31.111.134, 45ms,ae1-30g.ar1.ord1.us,0%

13, 69.31.111.10, 41ms,tge8-1.ar2.ord1.us.,0%

14, 68.232.176.32, 44ms,68.232.176.32.choop,0%
 
BTRY B 529th FA BN said:
Cause of buffering between the two NICS?
Click to expand...

There is a slight overhead on the driver (computer) side as it has to determine which NIC port to send the data based on which port is less utilized. Repeat that on the switch end as well. It's not noticeable in regular use but you can definitely measure it, it's usually in the single digit millisecond depending on the gear you use.
 
386DX said:
There is a slight overhead on the driver (computer) side as it has to determine which NIC port to send the data based on which port is less utilized. Repeat that on the switch end as well. It's not noticeable in regular use but you can definitely measure it, it's usually in the single digit millisecond depending on the gear you use.
Click to expand...

Actually, utilization has nothing to do with it. Each "conversation" is assigned to one NIC and stays on that NIC for its duration.

There are various hashing algorithms you can do: source and destination IP, source and destination MAC, source and destination port. Utilization is not one of them.

This is true for all forms of link aggregation: LACP, PAgP, forced aggregation, and driver-based ARP round-robin.

Link aggregation != load balancing.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top