Diagnosing Internet Connection Problems

Diagnosing Internet Connection Problems

Introduction

There are many things that can go wrong to interfere with a reliable internet connection.  Usually, one can systematically isolate and correct the problem without too much difficulty by knowing the wiring diagram of the network and which devices are affected by the problem.  I’d like to discuss a recent case that happened on my home network to help others who might have a similar problem.  The case involves a faulty streaming Blu-ray player which adversely affected the internet connection of every device on the network; even those not on the same branch of the wired network.

Symptoms of Problem

The problem manifested with the following symptoms:

• When booting a device (computer, etc.) or waking it from sleep it would take 20 to 50 seconds to establish an internet connection. Normally, a device should have established its connection by the time the device has booted or awakened.
• Internet devices would occasionally and randomly disconnect.
• The tested internet bandwidth was not affected.
• Routers used as wireless access points would intermittently fail to connect to devices.

Diagnosis – Phase 1

In diagnosing the problem, since the entire network was affected, I first looked for a global type problem doing the following:

• Updated the firmware in the Sonicwall.
• Reset the Sonicwall to factory defaults and reprogrammed it in case a faulty configuration file was causing the problem.
• Updated the firmware on one of the router/wireless access points.
• Checked the patch cables being used on one of the computers and found 2 of them to be bad (would only transmit at 100mbps).

I replaced the connectors on the 2 bad patch cables.  One was “fixed”, the other was discarded.  This got me to thinking that although the entire network was affected by the problem, could one bad device on the network adversely affect the entire network?  I did some internet research and found reports of bad network interface cards, switches, etc. affecting the entire network, not just the branch on which they were located.

Diagnosis – Phase 2

To find the rogue device I did the following:

1. Connected my laptop directly to the cable modem.  The internet connection came up instantly.  This verified that my laptop, modem, and internet service provider were all functioning normally.
2. Connected the modem to the Sonicwall (router) and my laptop to the Sonicwall. (everything else disconnected).   It worked fine.
3. Connected the Sonicwall to the computer room’s 16 port unmanaged switch with multiple devices and another router/wireless access point/switch connected to that.  I connected the laptop also to the 16 port unmanaged switch.  I left the rest of the house disconnected.  Everything worked fine.
4. Connected the cable going to the rest of the home network into the 16 port unmanaged switch.  The problem recurred.  So at this point, I knew everything in the computer room was fine and the problem was somewhere in the rest of the house.
5. A five-port switch in the basement acts as a hub to connect the Sonciwall to multiple switches in the rest of the house.  This was skipped initially due to access/convenience considerations.
6. The downstairs addition contains a router/wireless access point/switch which also connects to another switch in the upstairs addition.  I disconnected the downstairs switch from the system, but the problem persisted.  This meant the problem was not located in the house addition; neither downstairs or upstairs.
7. The TV room contains an 8 port unmanaged switch and a router/wireless access point/switch connected to it.  I disconnected the source cable from the 8 port unmanaged switch. Everything worked fine!  This located the problem to the TV room.
8. Unplugged all the devices from the 8 port unmanaged switch in the TV room and added them back one at a time.  Everything worked fine until I connected the Blu-ray player.  At that point, the problem recurred.  The rogue device was found!
9. Later I discovered the Blu-ray player had “crashed” and had to be rebooted.  I changed its internet connection from wired to wireless, and it did not adversely affect the network.  I have yet to test whether or not it was a defective wired network adapter in the Blu-ray player or just a side effect of its crashed state.

Discussion

So, the lesson to be learned is that the entire network can be affected by one bad device on it, regardless of its location on the network.

In further researching this issue I learned that using a “smart” switch or managed switch, could protect a network from this type of event.  Devices can sometimes cause network loopbacks or network storms that can utilize so much of the network resources that it can slow or stop the network.  The smart or managed switches can isolate the problem automatically, protecting the rest of the network.  This is something I plan to implement in the future.  I didn’t understand why my office computer vendors would recommend buying the more expensive managed switches over the cheaper unmanaged switches, which seemed to work just fine (until they didn’t).