Today I was called out to look at an unusual problem where the customer was able to visit some websites as usual, but not others. For example, going to ‘google.com’ worked, but trying to go to ‘facebook.com‘ or ‘national-lottery.co.uk‘ didn‘t. The web browser that I was using was Google Chrome, which said ‘This site can’t be reached’ and produced an ‘ERR_ADDRESS_UNREACHABLE’ message. However, in case there was a problem with Google Chrome, I tried using Mozilla Firefox instead. Unfortunately the result was similar, so it was not a browser-specific problem.
This was very strange, because normally the Internet either works or it doesn’t. I wondered what on Earth could cause it to work for some websites, while failing for others. At first I thought it might be a problem with the customer’s laptop, but when I tried accessing the Internet with a mobile telephone hooked onto the same Wi-Fi that the laptop was using, exactly the same thing happened. However, when using the mobile telephone’s 4G data connection, it was possible to access websites that were inaccessible through Wi-Fi. It seemed that the problem was being caused by the wireless broadband router, and that any device using its Wi-Fi was having the same problem.
When trying to diagnose Internet problems, there are some simple programs that can be used for troubleshooting. One of these is called ‘ping’, which works a bit like a submarine’s sonar system. It sends out a small packet of information to a specific place on the Internet then waits to see if it gets a reply, a bit like a submarine sends out a sound pulse and waits to see if it reflects off anything solid and bounces back. I opened a terminal window, often called a ‘command prompt’, and typed ‘ping www.national-lottery.co.uk’. Normally I would have got a response from the National Lottery website, but in this case I got nothing back. I wasn’t actually expecting a reply, but what I really wanted was to find out where ‘ping’ thought the National Lottery website was located on the Internet. The answer was ‘90.255.255.14’ and this looked wrong to me. The two ‘255’ numbers in the middle were not what I expected at all, so I reckoned the wireless broadband router was not looking up web addresses properly in all cases.
Although it’s a tired cliché in the computer industry, switching equipment off and on again often fixes problems, so I tried that using the wireless broadband router, but everything was just the same after the Internet connection came back. A router is just a specialised computer system, and you can usually log in to it by going to a special web address, which in this case was ‘192.168.1.1’, so I did that. After logging in I went into the advanced settings to see the location of the directory where the router normally looked up website names to convert them into Internet addresses. The answer was ‘90.255.255.90’.
When anyone using the wireless broadband router wants to go to a specific website, they simply start up a web browser and type in the website name, such as ‘www.national-lottery.co.uk’ for example. The router then goes to its standard directory location, which in this case is ‘90.255.255.90’, asks the directory located there to translate the name into an Internet address, and the request for the web page is sent there. In this case the address of ‘www.national-lottery.co.uk’ ought to have been returned as ‘194.246.78.103’, not ‘90.255.255.14’, so the problem was that the standard directory was not producing the correct Internet address.
Google just happen to provide their own directory facilities that anyone can choose to use instead of the one provided by their Internet service provider. Therefore, to solve the problem that we were having, I simply changed the router’s standard directory location from the default ‘90.255.255.90’ to ‘8.8.8.8’, where Google’s directory is located. Everything then started working as expected. Routers usually have a primary directory location and a secondary one to use if the primary one is not working. I therefore set the secondary direction location to ‘8.8.4.4’, which is Google’s secondary directory server.
It was possible that this problem was affecting other customers of the same Internet service provider, so I considered reporting it so the directory failure could be sorted out for everyone. However, I soon had second thoughts. I've tried reporting solutions to large companies before, but they’re usually high-handed and don’t want to know. They’re huge technology companies and I’m just an individual, so how can I know anything that they don’t? Big corporations are unhelpable, but my customer was happy, so I’d done what I was asked to do and that was enough.