Discussion Locked

PING interpretation

0 helpful
18 replies
Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.
In troubleshooting very slow response from our RFs I'm attempting to ping the client several times while they're experiencing a problem. The ping returns an approx round trip time of Average=400ms. This seems too long on a 10Mbps wireless network w/ 100Mbps backbone. What does this time tell me about the health of my network? Some of my better RF devices return a response time on average of 80ms. Could this indicate why my devices are so slow and timing out waiting for data from the host? Can someone interpret this for me?

18 replies

by Member 26122104

If these RF devices are able to step down their speed based on signal quality then they are running at a very low speed. Most likely though, there is interference that is effecting the signal quality causing either packet loss or retransmitted frames.

What types of devices? they should have counters that show these errors, as well as others.

HTH

by shanghai sam

We have error logs on the CISCO 350 Access Points that do look at packet loss. The logs are telling us that packet loss occurs. On a wireless network it is normal to have some packet loss. We are within normal range. The wireless devices are running too slow and do have the ability to step down. They stepped down to about 4Mbps from 10Mbps. The question is why? Our network has been gone through with a fine tooth comb. The ports on the switch are not logging excessive CRC/alignment errors, so therefore I know the CAT 5/fiber backbone is healthy. Thanks.

by ebob

The next step I would have a look at is "traceroute" ("tracert" in Windoze-speak) to the device. This is just to confirm that the "problem" is with the last hop to/from the device.

Assuming that it is the last hop, then it suggests that there is some "error condition" over this hop. (OK, so that's not news...)

It could be some form of interference. This is likely related to the location of the RF device, in relation to the AP. Perhaps the problem device is located too close to a fluorescent light fixture (or there's one in the "path") or some other EMR source.

I would be tempted to simply try relocating the trouble device(s) to see how that affects your results. If the problem goes away, then the source is "environmental". If thetrouble stays, then you might simply have one or more faulty devices.

Do these devices have logs that you can review? Are they showing lots of retries, or do they show that your devices are stepping down? How about your LAN devices? do they show lots of retires when talking to these RFs?

by shanghai sam

We placed in new routers at each end and actually tried a completely different connection circuit to the host with 100% bandwidth available. Still had problems. Thanks anyway.

by cavedweller

Try relocating the antennas. Sometimes a few inches or a foot can make a big difference with interference.

by shanghai sam

We have moved the antennas 3x already to completely different locations each time. Coverage actually degraded after doing this so we put them back according the original site survey results.

by akjacks

The RFs used in this network are 2.4 GHz Telxon 960SL PhaseIII portable transaction computers if that helps. The APs are CISCO 350s attached to yaggie antennas and 1 omni- directional. If interference is the problem, what average device would be inthe 2.4 GHz range for the 802.11b spec? Motorola radios, AM/FM radios, microwave ovens, ? Any other ideas?
Angela

by akjacks

Our site surveys tells us that signal strength is optimal for the current location of our antennas. We have several APs in our warehouse. During an AP handoff packet loss occurs. I discovered this by repeatedly pinging the client while they were moving around the warehouse. Cisco says that this should not happen, but it does. I would like to talk to a Walmart warehouse RF tech. I heard Walmart uses similar hardware and talks to an OS400 host via TN5250. At least they used to unless they recently moved to CE.

by ann777

Flourescent lighting can also be a problem. Try with the lights off and compare your results.

Electrical clocks and monitors, tvs, etc. all of them can cause interferance. Even solid cement walls with rebar inside can do it.

How many computers in your network?

by shanghai sam

The APs are located quite a safe distance from any RFI issues. But there is some type of interference from a distant source such as a microwave.

by stillatit

Looking at the rf issues, are the routers interfering with each other? If you take down all the routers but one, what happens to the ping time to the remaining devices? When you add back the rest, is there one which adds most significantly to ping times?

The long ping times are indicative of heavy traffic (including bad packets and retries). It may be that you have a bad remote unit which is "chattering." Try turning off all the remotes, and turning them back on in groups. If a certain groupseems to make things worse, find the bad unit in the group.

Looking at the handoff, this is supposed to happen smoothly because the routers can talk to each other. Check the manual -- this may be out-of-band communication, and may need to be direct. If they are using directional antennas, all pointed at a central box, they may not be able to talk quickly enough (or at all) to handle the hand-off.

Good luck.

by shanghai sam

We do not have have have heavy traffic from what we can see. Each AP has only 6 devices attached to it at any one time. Cisco states that you can you can 4x that easily. All the APs are fairly new and have been tested. All switches and AccessPoints have the latest firmware installed. All APs are tuned perfectly for our implementation according to CISCO. Thanks.

by akjacks

We have 100 wireless PTCs and 10 APs spread over 225,000 sq ft. We use directional antennas. We have tried testing the APs in groups and taking them off a switched network. There are 2 CISCO routers the data goes through before it hits the AS400 host. The APs and antennas are not located near flourescent lights.

by shmaltz

Looks to me like a window packet size problem. When you have different layer one devices IP will try to use the maximum allowed window size that the local Interface allows. When it comes to the bridge that has to convert it to a different size for the other interface. If the do not fragment flag is set it will require the bridge to drop the packet, When the client sees there is no reply it might try to resend the packet with a different size or if the do not fragment flag is not set it might take some time for the bridge to convert it.
You can test if this is the problem by using Ping and telling it the size and set the do not fragment flag. Here is how you do it in windows:
ping -l size (the size of the packet)
ping -f to set the flagping -l 1500 -f will send a packet of 1500 bytes and set the do not fragment flag.
Start out with a large size with flag set and then go down and see what happens. If you see that smaller packets come back faster the problem is because it can not take large packets so try decreasing the size of the packets the server sends out. This might be a problem since you don't want clients that are using the server on a network that can take bigger packets to communicate with smaller packets in that case you will have to use a router for the RFs.

by shmaltz

BTW Windows 2000 has default behaviors on how to handel this type of networks. I assume that you are not using Windows 2000 thats why I didn't mention it before. However if you want more info on how Windows 2000 handels this lookup the TCP/IP Core Networking Components book in the windows 2000 resource kit.

by shmaltz

The Ping commands I gave above works on Windows 2000 I don't know if it works on any other system.

This discussion is closed to new replies.