Hard to track down IP Conflict

[Inzoum]

I have a wireless network set up at home, and whenever I connect to it, Windows reports an IP conflict (usualy only once, on connect), but afterwards everything works perfectly... The thing is I don't believe there is an IP conflict since all the devices have manually assigned IP addresses within the same range and none of which are the same...

Let me explain the setup and steps I've tried to resolve this apparently unimportant (but annoying) issue:

Modem/Router:
Allied Data Copperjet 816 2P (Firmware 7.0, "Scarlet Box" provided by ISP)
- DHCP server enabled, Subnet value: 172.19.0.0, Subnet mask: 255.255.0.0
- Manually-assigned IP Address: 172.19.3.1, Subnet Mask: 255.255.0.0
- MAC ending with DB:8D

Wireless Access Point:
D-Link DWL-G700AP (Europe) (Firmware 2.10)
- Wired into router by ethernet cable (auto-selecting straight/crossover)
- Mode: Access Point, Channel 8
- Authentication: Open System
- SSID: Maison
- WEP 64-bit key
- DHCP server disabled
- Manually-assigned IP Address: 172.19.3.100, Subnet mask: 255.255.0.0
- MAC ending with C4:70

Wireless Repeater:
Belkin Wireless Universal Repeater F5D7132 (UK) (Firmware 5.01.11)
- Mode: Repeater, Channel 8
- Parent SSID: Maison
- Repeated Signal SSID*: MaisonR
- WEP 64-bit key (from Parent SSID)
- DHCP server disabled
- Manually-assigned IP Address: 172.19.3.101, Subnet mask: 255.255.0.0
- MAC ending with 45:74

* The Belkin F5D7132 is capable of repeating the parent Access Point's signal under a different SSID, to make it easier to select the which part of the infrastructure will associate with a device requesting a connection. Setting both the D-Link AP and Belkin repeater to the same SSID often led (in my case) to the PC connecting directly to the D-Link AP, bypassing the repeater entirely, resulting in a weak, unstable signal. In both cases the IP conflict is still present, so this most likely isn't the problem.

Desktop Computer
- Windows XP Service Pack 2
- Wireless PCI Card: D-Link AirPlus G DWL-G510 rev.C
- Auto-connects to MaisonR (Belkin Repeater)
- Manually-assigned IP Address: 172.19.3.5, Subnet mask: 255.255.0.0
- Default Gateway: 172.19.3.1
- DNS Server: 172.19.3.1
- MAC ending with EB:95

Playstation 3
- Auto-connects to Maison (D-Link Access Point)
- Manually-assigned IP Address: 172.19.3.10, Subnet mask: 255.255.0.0
- Default Gateway: 172.19.3.1
- DNS Server: 172.19.3.1
- MAC ending with (... I'd have to check)


The problem
- When the Windows XP PC connects to "MaisonR", it immediately pops a bubble (or alert window if in the Welcome screen) stating there is an IP conflict with another system in the network.
- Windows's System Events log reports: "The system detected an address conflict for IP address 172.19.3.5 with the system having network hardware address --:--:--:--:45:74. Network operations on this system may be disrupted as a result." (This is the Belkin repeater's MAC, as you can see: both devices have different IPs in the same subnet, where's the conflict?).
- Everything is perfectly fine afterwards, the alert message only pops on connecting, not a big problem, but a real annoyance, and doesn't look very professional

Things I've tried

Switching both Belkin/D-Link devices around (using the Belkin as Access Point and D-Link as Repeater)
- IP Conflict remains, but the conflicting device in the logs now point to the D-Link's MAC address
- The D-Link device is unable to repeat the parent signal under a different SSID, it has to use the same.
- The D-Link device will disconnect and re-connect every 3 minutes if there is no network activity when setup in Repeater mode, which causes my PC to re-connect to the "Maison" SSID, often connecting to the Access Point behind (with a weak signal) while the D-Link repeater is busy re-associating
- Since the D-Link keeps re-associating every 3 minutes if there is no activity, it dies overnight due to excess erros and needs to be rebooted (which is a mess when the PC tries to auto-connect to the far-off access point in the meantime)
- All in all: VERY BAD SETUP, the D-Link was meant to be an AP, and Belkin meant to be a repeater, not the other way around (even though they both devices have both modes, the options available for each are clearly different)

Observations
- The conflicting device is always the repeater's mac address, no matter which device is set as the repeater
- The Playstation 3 never reports any problem, whether I connect it to the repeater's SSID, or the Access Point's SSID (the AP is closer, but both have a strong enough signal relative to the PS3)
- I never had any IP conflicts in my old appartment where a repeater wasn't necessary, I only had the D-Link Access Point then and everything was perfect
- I regularly check for intruders, and I've never seen any connecting to either part of the wireless infrastructure

My feeling is that this is really just a bug with Windows XP and connecting to a repeater, but it's annoying nonetheless.

Any kind of help is appreciated.

Thanks!

Luc René de Cotret


EDIT:

I take back what I said about the conflict occuring only on connect, I just had a nasty case just now, had to re-start the repeater and re-connect... I browsed other forums about this, and from what I've found, this seems to be a general problem with Windows XP, Wireless Repeaters and Gratuitous ARP... now how to solve it is another thing

[johnwill]

I avoid wireless repeaters like the plague, just another piece of equipment to tinker with and configure.


Hawking Tech has a number of products that will help you increase your wireless range. The root page is Hawking Hi-Gain™ WiFi Range Extending Products.

Some of the more interesting products are this Hawking [HSB2] Hi-Gain WiFi Signal Booster, which can be used on either end of a wireless connection to boost the signal power.

Another way to increase your signal strength is by the use of hi-gain antennas. You can choose from omni-directional or directional models, here are a some examples.

Hawking [HAI7SIP] Hi-Gain 7dBi Omni-Directional Antenna

Hawking [HAI15SC] Hi-Gain 15dBi Corner Antenna

[HAO14SD] Outdoor Hi-Gain 14dBi Directional Antenna Kit

[Inzoum]

Thank you for your suggestions...

For reference, I think I understand what may be happening... The Playstation 3 gave me a hint, actually...

I have TVersity set up on my PC for sharing media over to the PS3. When the Media Server is detected by the PS3 (sadly it seems that with this whole repeater thing, I can only wait for the Media Server to announce itself by broadcast as doing a search from the PS3 yields no results), it is listed as being at IP 172.19.3.5 with MAC address ending with 45:74 (Wireless Repeater's MAC, not the PC's).

This leads me to believe that devices on the other side of the Wireless Repeater see all packets coming from 172.19.3.5 as being sent from machine 45:74. When other machines send packets to 172.19.3.5 at 45:74, the Wireless Repeater knows these packets truly belong to the PC behind, with the EB:95 MAC address. Packets are sent back and forth properly, but there is some confusion regarding Address Resolution.

When a Windows machine connects to any network, it sends a Gratuitous ARP broadcast to resolve its own IP address (and make sure there is no conflict with other network machines), the problem with wireless repeaters in my case is that other machines on the network see 172.19.3.5 as belonging to 45:74, and not EB:95, so my PC is being told by the network that my Repeater is already using my IP address. A fight ensues.

Why this particular case hasn't been thought out properly in Wireless Repeaters is completely beyond me. What were they thinking? Don't they test their products? Didn't anyone at D-Link or Belkin or Microsoft notice that all hell broke loose in the ARP protocol when a Windows machine connects to a Wireless Repeater? Would it be that difficult to program a little packet analysis program in the Repeater that checks for Gratuitous ARP packets, and cheats a little to reassure Windows? Would it be that difficult to program something in Windows to recognize Wireless Repeaters and consider the MAC address analogy that ensues (and not worry about it)?

Really, I'm completely dumbfounded. Flabbergasted. It really seems like a good, powerful idea that comes crumbling down because of a tiny detail no one ever thought about on either side of the deal. David vs Goliath. And they went on and set everything in stone with standards no one's supposed to break, sealing the broken nature of Wireless Repeaters for all eternity. Amen. With all this said, I hope others can learn from my experience.

All in all, I'll stick with my current setup until I can gather the courage to do some wire-work. The only reason I have a repeater set up is because my phone line (and therefore Modem/Router and Access Point) is at one end of the apartment and my PC is at the other end, the repeater being in the hallway, smack in the middle. It's possible to connect to the repeater from anywhere in the apartment with a good signal quality, so if I extend an ethernet cable to the hallway and move my Access Point there, I won't even need a Repeater anymore since the whole apartment will be decently covered. What's making me cringe is that I'll have to go around a door frame and drill through a wall to get the cable there (and since I'm a perfectionist and don't want a behemoth-class hole in my wall, I'll have to borrow someone's tool to crump the ethernet plug on the cable myself after passing it plug-less through a discrete, tight opening)

Cheers!

[johnwill]

That's why I stick to either wired connections or extending the range of the source or destination. I've tangled with repeaters a number of times, and my batting average is less than .500 in getting a satisfactory result.

[fuller_paul]

Did you figure out any solution to alleviate the problem? I am having the exact same issue with a very similar configuration (DLink Wireless Router and Belkin Repeater). What's bizarre is that my wife's XP home laptop is not getting the same issue as my children's XP home PC.