Sign in with
Sign up | Sign in
Your question

How We Test Hardware Routers 2006

Last response: in Toms Network
Share
May 23, 2006 6:48:13 PM

We've changed how we test hardware routers. Questions and comments on the new procedure go here.
June 8, 2006 11:00:32 PM

Could you provide the IP addresses being used on both the Lan and Wan sides of the router for this test. I am assuming that you are using a static IP on the computer connected to the Wan port. Traffic going from Lan to Wan is ok but going from Wan to Lan never makes it through. Keep getting error no endpoint at this address. Could you help out with this one.
June 11, 2006 3:25:58 PM

Quote:
Could you provide the IP addresses being used on both the Lan and Wan sides of the router for this test. I am assuming that you are using a static IP on the computer connected to the Wan port. Traffic going from Lan to Wan is ok but going from Wan to Lan never makes it through. Keep getting error no endpoint at this address. Could you help out with this one.

The IP addresses used depend on the router and network setup. I actually connect the router WAN port to my LAN's switch and let it grab an IP address from my DHCP server. The WAN test computer is also connected to the switch.

Since WAN to LAN traffic is rejected by NAT routers by default, you must put the LAN machine in DMZ or open the ports used by your specific test application. In the case of Qcheck/IxChariot the ports are TCP/UDP 10113-10117. You also must use the router's WAN port IP address as Endpoint 2, since that is the only address seen on the WAN side of the router.

Even then, for some routers using NAT+SPI, the test might run, but you won't get results returned due to limitations in Qcheck/IxChariot. The only way around this is to try a different application like iperf/jperf

http://dast.nlanr.net/Projects/Iperf/
Related resources
June 17, 2006 1:07:31 PM

Quote:
We've changed how we test hardware routers. Questions and comments on the new procedure go here.


I am responsible of network system in a 600 m2 general store (1 floor down,2 floors up).So I need very long ranged access points to minimize the number of repeaters.I searched many sites to find out which router has the best range and at the end only your chart and related reviews helped me to choose. There is one thing I want to suggest that range performance test may differ 802.11b only, 802.11g only and mixed modes, for example hand terminals that we use communicates via 802.11b mode so 802.11b mode range performance is the most important to me.One more thing , I wonder the performance of Asus
WL-566gM router.
June 20, 2006 7:23:44 AM

Quote:
Since WAN to LAN traffic is rejected by NAT routers by default, you must put the LAN machine in DMZ or open the ports used by your specific test application. In the case of Qcheck/IxChariot the ports are TCP/UDP 10113-10117. You also must use the router's WAN port IP address as Endpoint 2, since that is the only address seen on the WAN side of the router.


What do you mean by "Endpoint 2"?
I configured a network around a WRT54Gv5 in the following manner:
LAN comp. (192.168.1.101/24 DG:192.168.1.1) connected to the LAN port on the router. WAN comp. (10.0.0.1/24 DG:10.0.0.254) connected to the WAN port of the router. i gave the router a static IP of 10.0.0.254/24 with a bogus DG and DNS (10.0.0.250).
I've configured the router to use 192.168.1.101 az DMZ.

now, communication between LAN to WAN works flawlessly. I managed to download a 550MB file from an FTP server on the WAN comp. using hacked flashget which divided the file into 30 threads..

on the other hand, communication WAN 2 LAN was unsuccesful..

After all said and done (sorry for the length ;)  ) , why, oh, why wont the WAN comp communicate with my LAN?
pinging 192.168.1.101 resoltet "request timed out".

The LAN comp is an XP SP2 with Firewall disabled.

I thought putting the LAN comp in DMZ should enable it. I allso disabled the "Block Anonymous Internet Requests" option in the router.. didn't help..

is there something i'm missing here?

oh.. wait a minute...
i think i got it :oops: 
since the router is the only thing that the WAN sees, pinging 10.0.0.1 is actually pinging the computer behind it, to whom all ports are forwarded... right?

:oops: :oops: :oops: :oops: :oops: 

Never mind... i'll post it just so other dumbs like me will know ;) 

PS.
in the article you wrote:
LAN to WAN - Data flows from the WAN-side IxChariot endpoint to the LAN-side endpoint. This is a test of router upload speed.

i guess you've meant "from the LAN-side" :-)
June 20, 2006 12:48:01 PM

Quote:
since the router is the only thing that the WAN sees, pinging 10.0.0.1 is actually pinging the computer behind it, to whom all ports are forwarded... right?

That's correct. When running the WAN-LAN test, endpoint 2 in IxChariot or Qcheck must be set to the router's WAN IP address.

Quote:
PS.
in the article you wrote:
LAN to WAN - Data flows from the WAN-side IxChariot endpoint to the LAN-side endpoint. This is a test of router upload speed.

i guess you've meant "from the LAN-side" :-)

Yes, you're correct. Sorry for the error. It's been corrected.
June 24, 2006 6:34:21 PM

Personally I would like to see an AMD based system (non-64bit) mixed in the test computers. Actually for both wireless NICs and routers.

I have searched on the web many times and haven't yet found a review of products tested on an AMD system. I also searched the forums here for AMD and it came back with no results.

On the Circuit City site I did see a user review that had an AMD system that made a general statement that Netgear products and AMDs don't get along. I don't know if it was opinion or fact.

Anyway, just my $.02 on what I would like to see in the review.

betty
June 25, 2006 1:46:56 PM

Quote:
Personally I would like to see an AMD based system (non-64bit) mixed in the test computers. Actually for both wireless NICs and routers.

I have an Athlon 64 3000+ system that I use from time to time for testing and have never had a problem.

Since all the tests are done with TCP/IP network traffic on Windows XP, I can't think of a test where Intel vs. AMD would make a difference.
June 25, 2006 5:58:46 PM

Thank you for the reply Tim-

I had read seen the AMD 64 mentioned which is why I put the (non 64bit) statement in. Just so you'd know I really had tried to find it before asking :wink:

I had read reviews all over the net, and many of the Pro ones on the wireless routers and nics varied so much, I started reading consumer reviews. Those vary so much also on all the products that I still am using my old Linksys BEFSR41 router and the internal nics on the 2 systems.

Another reason, besides not truly knowing if AMD makes a difference or not is I had posted on a different networking forum about what would be a good wireless set-up and was told by a few that since my cable connection is 7Mbps that it would be a waste of money to buy any of the 802.11g and faster hardware and just stick with the Linksys.

Anyway, I'm sorry if the AMD issue is really not related to how the networking hardware functions, however hopefully you'll understand my confusion and why I thought it was an issue along with answers about my cable speed.

One other question though that hopefully is related to router/nic function is that my cable is Roadrunner and I also have the digital telephone so my modem is a two part one. One side is the internet connection, the other side is the telephone interface. Will that affect network speed and function? Has this type of modem been tested with the wireless hardware?

I won't post in this thread again, since it is for a specific purpose.

Sorry,

betty
June 25, 2006 8:45:36 PM

Quote:
I had read seen the AMD 64 mentioned which is why I put the (non 64bit) statement in. Just so you'd know I really had tried to find it before asking <snip>

You're welcome to post in this or any other thread any time. :) 

The "best" product for you depends on a number of factors. For the routing part of a wireless router, it's true that you're not buying yourself any advantage by purchasing a router with faster wired routing speed than your Internet connection.

For the wireless portion, you can take advantage of speeds higher than your internet connection for tranfers between computers on your LAN. Since these transfers don't touch the Internet, they aren't limited by the speed of your connection.

Wireless speeds higher than your Internet connection can also help because they give you a higher starting point for the speed vs. range curve that all wireless routers have. But if you have a lot of neighbors close by with wireless LANs, you could find yourself with low and/or erratic speed due to interference from those neighbors.

But if you have no or little interference, aren't trying to cover too large an area with wireless, and are going to be doing mostly web browsing and email while wireless, an 802.11g wireless router could serve you just fine.
June 26, 2006 2:44:11 AM

Tim said in part:
Quote:
For the wireless portion, you can take advantage of speeds higher than your internet connection for tranfers between computers on your LAN. Since these transfers don't touch the Internet, they aren't limited by the speed of your connection.


I had wondered in the past if since one of the systems has a built in Intel gigabit nic if it would be worth buying a 10/100/1000 pci card for mine (vs. the built in 10/100 conenction) and getting a gigabit switch to connect them for file transfers between systems and then getting wireless cards and router for internet use only. I don't even know if this is possible or feasible.

then Tim said:
Quote:

But if you have no or little interference, aren't trying to cover too large an area with wireless, and are going to be doing mostly web browsing and email while wireless, an 802.11g wireless router could serve you just fine.


A couple of months ago I did decide to try the wireless network (which is why I'm still with the wired Linksys). I bought the Belkin Pre-N router and matching wireless card. Ended up having to call Belkin tech support as no matter what we tried, could not get an internet or pc to pc conenction. The Belkin tech guy after going through all the steps said it must be faulty to take it back. So I exchanged the card and router for a Linksys wireless G and card. Same thing happened with the internet although we did connect pc to pc. So I just returned those and didn't buy anything else.

With the Linksys though we did see that the neighbor on one side has a wireless connection. That was the only connection we saw. It said the strength was around 57%.

Would that interfere?

I assume from your reply that anything above a G would be a waste of money at this time? Basically then would any of the current G routers (except the Linksys one - I read the article!) be ok?

Thank You again,

Betty
!