Testing Latency Stability

[tential]

I'm wondering how I can test latency stability. I want to get a powerline adapter and I plan on it. But I want to test it to see how it compares to my wired ethernet and my wireless. Any good way to get graphs so I can compare?

 

[Bubbaleone]

I'm not sure what you mean by "stability". If the TCP/IP stack and network adapter settings are all good; latency becomes a function of what type IP service you have ( dial-up, DSL, Broadband, Satellite), how many hops to the CO, how many hops from the CO to the URL you want, and how much congestion there is over the backbone (which varies constantly in any 24 hour period) at the time of day you're connecting . There's no way of detemining latency without actually having the powerline adapter connected at your geographical location.


.

 

[tential]

I'm not sure what you mean by "stability". If the TCP/IP stack and network adapter settings are all good; latency becomes a function of what type IP service you have ( dial-up, DSL, Broadband, Satellite), how many hops to the CO, how many hops from the CO to the URL you want, and how much congestion there is over the backbone (which varies constantly in any 24 hour period) at the time of day you're connecting . There's no way of detemining latency without actually having the powerline adapter connected at your geographical location.


.

I want to test how much packet loss I'd get from UDP? I think that's what's important for gaming right?

If I set up the powerline adapter and test UDP packet loss wouldn't that show how much packet loss the Powerline adapters are adding in?

 

[Bubbaleone]

I'm not an expert, but User Datagram Protocol packet loss testing requires specialized software such as NetScanTools® Pro, and extensive knowledge of networking protocols. It's not like running a TCP ping command from the command prompt. You may find this article a useful resource for what to realistically expect from powerline adapters: SmallNetBuilder's Powerline FAQ

 

[tential]

The problem is I'm trying to use it for gaming so I need something to measure whether I'm getting spikes because of the powerline or because of just spikes in the internet.

It mentions Iperf, and in another thread on another forum they mention that as well. Would that be good to test UDP packet loss? I'm very unfamiliar with the terms, a lot less familiar than you guys are probably especially since I've barely used computers other than for word processing and web browsing these past 5 years. Gaming has taken a backseat and I dunno much about new technologies or how far old technologies has progressed. I'm playing catch up so anything will help.

 

[BTRY B 529th FA BN]

Your easiest bet is to download Ping Plotter and see which hop gets a report of Packet Loss. e.g. the game server I play at there is a hop that is F*cking horrible @ us.nlayer.net. Consistently dropping packets. It's address is 69.31.111.243

 

[XavierMace]

Download a 3rd party utility to do a UDP ping (IE nmap.org) and ping from your computer to a device on the other side of the powerline but before your internet connection. IE your router.

In general, powerline adapters are in between wireless and traditional wired networks in performance and reliability. Performance varies depending on house wiring and the quality of the equipment.

I also think you might need to clarify what's causing you to think this is a major concern. It's got to be a pretty crappy connection to cause issues gaming.

 

[XavierMace]

Your easiest bet is to download Ping Plotter and see which hop gets a report of Packet Loss. e.g. the game server I play at there is a hop that is F*cking horrible @ us.nlayer.net. Consistently dropping packets. It's address is 69.31.111.243

Ping Plotter doesn't test UDP communication.

 

[BTRY B 529th FA BN]

Ping Plotter doesn't test UDP communication.

Ok?

 

[tential]

Download a 3rd party utility to do a UDP ping (IE nmap.org) and ping from your computer to a device on the other side of the powerline but before your internet connection. IE your router.

In general, powerline adapters are in between wireless and traditional wired networks in performance and reliability. Performance varies depending on house wiring and the quality of the equipment.

I also think you might need to clarify what's causing you to think this is a major concern. It's got to be a pretty crappy connection to cause issues gaming.

It's not pretty crappy, I just have never used powerline before. I'm used to wired connections. I don't trust wireless due to generally getting spikes every once in awhile. Even one spike during a game just irritates me.
I want to be able to compare the difference between ethernet, my current 10 year old powerline adapter, and the new powerline adapter I plan to buy by the end of the week. Depending on performance, I'll know what to do.

Similar to how you wouldn't buy a new GFX card without knowing the performance, I don't want to invest in using a technology I haven't used unless it increases the performance.

I'll test out these programs you guys have listed so far:
nmap.org
iperf
Ping Plotter

Anything else I should use?

 

[XavierMace]

Ok?

He's asking how to test UDP performance. All Ping Plotter is doing is giving you a pretty GUI to standard ping/tracert results.

It's not pretty crappy, I just have never used powerline before. I'm used to wired connections. I don't trust wireless due to generally getting spikes every once in awhile. Even one spike during a game just irritates me.

Then I'd prepare for irritation. Not trying to be a smart ass, but you've got no control over what happens after the connection leaves your residence which is probably where the spikes are happening.

The other reason I ask is because I've played MMO's tethered off my cell phone on 3G. For you to get enough packet loss on a LAN to cause an issue, it would have to be a LOT of packet loss.

 

[BTRY B 529th FA BN]

He's asking how to test UDP performance. All Ping Plotter is doing is giving you a pretty GUI to standard ping/tracert results.

In my experience with Unreal Tournament III, TCP packets are still involved. Although Ping Plotter isn't entirely verbose about what packets are being dropped it will indicate a hop that is having trouble and thus give an indication of where trouble is being had with packets. Whether or not it's UDP packets, it's still an indicator. And yes, when that server I mentioned above, the n.layer.net server, was having 60% packet loss for seconds to minutes at a time, it was noticeable even though the game is predominately UDP

EDIT: you can watch packets via Wireshark

 

[XavierMace]

I'm not arguing a ping not telling you anything. I'm just pointing out it doesn't check UDP traffic. If they are dropping packets, yeah, you are going to see that on a ping. But if a firewall is blocking or throttling UDP traffic, you aren't going to see that on a ping.

 

[BTRY B 529th FA BN]

Cool

 

[ch33zw1z]

http://manpages.ubuntu.com/manpages/lucid/man1/iperf.1.html

iperf manual. option -u will test UDP instead of TCP

Just setup iperf on a machine on both sides of the powerline, test away.

If all you want to test is packet loss, testing of TCP should suffice. If the powerline connection is dropping packets, testing TCP is the easiest way to tell.

 

[tential]

http://manpages.ubuntu.com/manpages/lucid/man1/iperf.1.html

iperf manual. option -u will test UDP instead of TCP

Just setup iperf on a machine on both sides of the powerline, test away.

If all you want to test is packet loss, testing of TCP should suffice. If the powerline connection is dropping packets, testing TCP is the easiest way to tell.

Only reason I'm worried is because Anandtech's 2011 review of Powerline Adapters, Anandtech felt UDP packetloss was so high they didn't even include the numbers. I know a lot of people still use it for gaming, but like everyone says, it's dependent on your house. So I just want to check, make sure it's not my powerline adapter that causes lag sometimes when I play halo 4. In particular I just want to test the new Power Line Adapter I'm purchasing as well since I don't want to wait a week of playing online games then feel I don't like it and try to return it then. I'd rather purchase, test, then game for a day or two and if I don't like it return it. I do better with numbers rather than judging off experience.