A lot of people talk about lag. The purpose of this forum post is to provide a little education on what lag is, and what you can do to learn what or where the root cause is when you experience it.

Without further study or investigation, LAG is a term used to define the perceived slowness you experience when you attach to a website, or in our case GamingPeak.

First of all, it's important to note that it's NOT always correct to blame GamingPeak for your connection "slowness" or lag. In almost all cases, the lag is being caused by factors related to how your requests for data from GP travels to the PC you are using. It's really all about the journey - and it's a pretty intricate journey sometimes, for some people (especially overseas players).

It might be good to review how the internet works (in very simplistic terms). First a little background:

The data you get from websites, and send to websites (ie, "exchange") between your PC (called a client) and any website (called a server) is known, in simple terms, as data packets, or "packets" for short.

These packets of data do NOT ever travel directly from your PC to the website, or server in one direct connection. The web isn't designed this way (and for a good reason, but more on that in another educational piece).

In our case, data packets travel from server to server along the way. Each server's job is to move the data along - kind of like a shepherd - from your PC and to the final destination - the website you want, and then move the data back from the website to you. Each of these servers represent a "hop" in the pathway your data takes. Hops then, represent the various locations your packets see or "touch" as they transit (or site see) from your PC to your final destination.

For example, when the data leaves your PC, it may first travel to your router (1st hop), then your highspeed modem (2nd hop) then to your local ISP server in your town, (3rd hop), then to that ISP's bigger server in the largest town or city in your area (4th hop), etc. It's going to eventually gets on the backbone of the USA internet, and that's similar to the interstate highway system in the US. It's also similar to the big arteries and veins in your body. Capillaries first turn in to small vessels, which get bigger and bigger as they get closer to the heart.

The aorta is called a "big pipe". Same is true for the internet backbone.

Nature does provide good models for some of this. We take from nature in many things we humans build.

In another example, think of this the same way you think of how you travel in your car when you want to go from your house in Newark NJ to San Francisco, CA (that's a cross country trip). You first leave your driveway, then get on your street. You then get on a little bigger street (more traffic), then maybe a 2 lane state route (even more traffic), then finally get on the 3 lane interstate and drive....and drive....and drive.

Luckily, your packets go at near the speed of light, but - the point is they kind of act in a similar fashion to how you would go in your car (as a rough analogy).

Each time your data packets hit a new server, that server is suppose to move them on their path across the country. This server can also report the time it takes to respond to your request to move your packets along to their destination. If the server is busy, or there is bad weather, or its got some other issues - that one server will be less responsive to you, and add more time than it should to your request for service from it.

When internet servers get slow, time goes up for what your doing and you experience that. You call it lag. You're being subjective, but you're right. Now let's be objective and see where that lag comes from.

It's pretty cool that you can see the path of your packets, and each server along the way will record how long it takes on each step of their journey. Before we take a look, note the following important things so you'll understand what you see:

1) a millisecond is 1/1000 of a second, and abbreviated as "ms". So 500 milliseconds is 500/1000 of a second, or simply 1/2 a second. That's a lot of time in the internet world. You don't want to see servers taking 500 milliseconds to handle your packets.

2) 50 milliseconds is a realistic, and a nominal value for each server to handle your data packets. In the world of milliseconds, ALL SMALLER NUMBERS are BETTER for you, always.

3) If you see a * in this evaluation, the server didn't respond in a reasonable time. A row of 3 stars is trouble, and the server has "timed out" and isn't responding to your request. Not good when this happens.

So how do you get this analysis for yourself? Easy - very easy in fact.

There is a tool on your XP or Vista PC called TRACERT (short name for what the program does - it TRACE ROUTE of your data packets). You can read about TRACERT on the web by googling it. If your PC doesn't have this tool, you can download a free one from a reliable website. Once you have it, you can try this:

1) Go to the start menu
2) Choose "RUN"
3) Type "cmd" or "command"
4) When the command box opens (aka a DOS box), type this: tracert gamingpeak.com.

This will trace the route of the packets from your PC to gamingpeak.com, and give you an analysis.

What's this look like? Take a look here. I've got an example which demonstrates this:

________________________________________
Microsoft Windows XP [Version 5.1.2600]
(C) Copyright 1985-2001 Microsoft Corp.

C:Documents and SettingsX4XD>tracert gamingpeak.com

Tracing route to gamingpeak.com [209.20.85.250]
over a maximum of 30 hops:

1 10 ms 1 ms 1 ms 192.168.155.102
2 1 ms 1 ms 1 ms 84.34.51.2
3 7 ms 7 ms 8 ms gw-44-216-ur01.43d.wa.seattle.comcast.net [67.85.160.129]
4 9 ms 9 ms 9 ms te-23-334-ar02.wa.livingstn.comcast.net [67.87.144.241]
5 13 ms 11 ms 11 ms 67.89.162.70
6 17 ms 15 ms 15 ms pos-1-3-0-0-cr01.chicago.il.ibone.comcast.net [69.186.90.157]
7 14 ms 15 ms 19 ms te2-2.mps01.jfk05.atlas.cogentco.com [154.54.11.209]
8 15 ms 15 ms 15 ms te2-4.ccr01.jfk02.atlas.cogentco.com [154.54.6.49]
9 21 ms 20 ms 20 ms te9-8.mpd01.bos01.atlas.cogentco.com [154.54.25.242]
10 42 ms 41 ms 42 ms te8-8.mpd01.ord01.atlas.cogentco.com [154.54.24.54]
11 42 ms 41 ms 43 ms gi12-0-0.core01.ord01.atlas.cogentco.com [154.54.1.97]
12 47 ms 47 ms 47 ms po14-0.core01.stl03.atlas.cogentco.com [154.54.3.66]
13 47 ms 49 ms 47 ms vl3802.na21.b016110-1.stl03.atlas.cogentco.com [66.28.5.246]
14 47 ms 47 ms 51 ms 38.104.162.22
15 47 ms 49 ms 52 ms 209.20.79.227
16 49 ms 48 ms 47 ms 209-20-85-250.slicehost.net [209.20.85.250]

Trace complete.

C:Documents and SettingsX4XD>

___________________________________________

Let's analyze this together:

a) There are 16 "hops" that my packets need to take to get from my PC to GamingPeak.com
b) My times for each hop are excellent!! All are less than 50ms, for the most part. No 3 digit numbers at all. This is like having all green lights ahead of you! I've got no lag here! I've got each server in the chain responding very quickly to my requests.
c) You can see the packets leave my house (in Washinton), then go to Chicago, and then get on the backbone, which is provided by Cogentco. Cogentco delivers the packets to Slicehost at HOP 15, which is the ISP that GamingPeak uses.
d) The final destination at hop 16, is what you call "GamingPeak". That's our IP address and, our server.

Pretty simple huh?

Now imagine you're experiencing moderate lag. How would this look? Let's see another example:

________________________________________
Microsoft Windows XP [Version 5.1.2600]
(C) Copyright 1985-2001 Microsoft Corp.

C:Documents and SettingsX4XD>tracert gamingpeak.com

Tracing route to gamingpeak.com [209.20.85.250]
over a maximum of 30 hops:

1 10 ms 1 ms 1 ms 192.168.155.102
2 1 ms 1 ms 1 ms 84.34.51.2
3 7 ms 7 ms 8 ms gw-44-216-ur01.43d.wa.seattle.comcast.net [67.85.160.129]
4 69 ms 69 ms 79 ms te-23-334-ar02.wa.livingstn.comcast.net [67.87.144.241]
5 113 ms 111 ms 111 ms 67.89.162.70
6 170 ms 150 ms 150 ms pos-1-3-0-0-cr01.chicago.il.ibone.comcast.net [69.186.90.157]
7 14 ms 15 ms 19 ms te2-2.mps01.jfk05.atlas.cogentco.com [154.54.11.209]
8 15 ms 15 ms 15 ms te2-4.ccr01.jfk02.atlas.cogentco.com [154.54.6.49]
9 21 ms 20 ms 20 ms te9-8.mpd01.bos01.atlas.cogentco.com [154.54.25.242]
10 42 ms 41 ms 42 ms te8-8.mpd01.ord01.atlas.cogentco.com [154.54.24.54]
11 42 ms 41 ms 43 ms gi12-0-0.core01.ord01.atlas.cogentco.com [154.54.1.97]
12 47 ms 47 ms 47 ms po14-0.core01.stl03.atlas.cogentco.com [154.54.3.66]
13 247 ms 249 ms 247 ms vl3802.na21.b016110-1.stl03.atlas.cogentco.com [66.28.5.246]
14 * * 300 ms 38.104.162.22
15 47 ms 49 ms 52 ms 209.20.79.227
16 49 ms 48 ms 47 ms 209-20-85-250.slicehost.net [209.20.85.250]

Trace complete.

C:Documents and SettingsX4XD>

______________________________________________

Let's analyze this and study where there are potential problems:

A) Hop 4 isn't bad, but Hops 5 and 6 are taking more time than they should. Remember, if each hop takes 1/10 of 1 second longer than it should, those 10th's will add up. 10 of them will add 1 second of time to your overall experience. That's a long time when you're waiting.

B) Hop 13 is a disaster. Something's wrong here, this hop is adding about 2/10 of a second to the lag of your packets. Not good.

C) Hop 14 has timed out in 2 out of 3 attempts. On the last attempt, roughly 3/10 of a second (300 milliseconds) was expended for it to respond to moving your packets. That's too long.

OK, so what have we learned?

1) If, when you have lag, you study your own situation first, you'll be in a much better position to KNOW where the problem is.

2) Run TRACERT a few times, and note where the block or lag seems to be. Look for servers or hops that consistently display higher numbers. As we said initially, lag can be caused anywhere along the data path your packets take. Sometimes in multiple places.

3) In my examples, my packets take 16 hops to get to GP. Your own situation will vary. Typically, fewer hops can mean an inherently faster conceptual trip, but not always. TRACERT will show 30 hops, so if you have more than that, ..... it's a long trip for you :) Think of this in this way: if you live right next to the entrance ramp to an Interstate highway, your commute to your destination will take less time than if you have to travel a bunch of side-streets and backroads to get to that same interstate entrance ramp.

4) If you see high response times (ie HIGH NUMBERS) in the last 2 hops of the journey, then it IS a problem with GamingPeak's ISP or GamingPeak's server. If that's the case, write to support@gamingpeak.com and they'll study it.

Hope this helps. 

Time2Change asks:

i ran a tracert to several different sites and at hop 2 it times out every time. I am on a wireless router and that is hop 1 which takes just 1 ms. when it gets to hop 2 it times out and i have no clue what hop 2 is because when it times out it doesnt give you the information. could hop 2 possibly be the cable modem? if so would a new modem fix this (im on comcast cable internet)? or am i just stuck with this situation for life? 

Short answer: Nothing is wrong.

Longer answer and education... start reading:

This happens a lot. It happens to me, as well - and some people have "sneaked" a few tracert outputs to me for private readings, and their hop 2 times out as well. Here's why:

If you have a router AND a broadband Modem, HOP 1 is your ROUTER (which provides your wireless and wireline connections) and the broadband MODEM is HOP 2.

Hint: Broadband is a collective term used for noting any high speed network connections - and can be DSL or cable, WIFI, etc. Narrowband is a fancy name for "phone dial up".

Broadband is NOT a term used to describe the wireless you use connecting your PC to your ISP modem or router. That's high speed already. Wireless-G is a 54Mb connection speed (way above what what any ISP delivers to your house).

Broadband is the term used to define what your ISP (aka Internet provider) gives YOU for service speed from the street - through the cable or optical fiber coming into your house. Comcast, Verizon are examples of broadband providers. Netzero is an example of a narrowband provider.

Frequently, ISPs will set their broadband MODEMS up (remember the modem? - that's HOP 2) so that they do not (and will not) respond to pings or ICMP/UDP packets. ICMP/UDP are protocols that support the packet types used for doing traceroutes. Doing that is a security thing, and a good idea :)

You can look up what a ping is, or what ICMP/UDP packet types are at your leisure. But here's a short version:

The UDP and ICMP protocols are part of the core protocols of the Internet Protocol Suite, which is the way everything talks on the internet.

They are chiefly used by networked computers' operating systems. ICMP for example, to send error messages - indicating, for instance, that a requested service is not available or that a host or router could not be reached.

UDP stands for User Datagram Protocol and ICMP stands for Internet Control Message Protocol.

In summary:  Broadband is a measure of speed.  Wireless is a type of connection to obtain that speed.

Anyway....

Since the modem isn't designed or permitted by the ISP to respond to those packet trace types, all it can do is show you it doesn't understand that, and that means it LOOKS to you like it times out.

If you don't know German, and I start talking to you in German - you're going to timeout and go "huh?" too... :)

But in fact, it's moot. The modem may not itself understand those packet types, but what it DOES do is (importantly) pass them on anyway for you, all the time it also shows you it didn't (or can't) respond for that HOP (which is - HOP 2).

The rest of the journey for those packets is fine and more representative of what's actually going on.

Basically NOTHING is most likely wrong at either HOP 1 or HOP 2 if you see a timeout like this.

Thanks for the information, Sundancer.  It's very interesting and makes some sense.

                                Sincerely,        Acer