I sometimes use a free Wifi service to get access to the internet.
Like most/all providers of services like this, this service employs a captive portal. So if you try to make a HTTP request (request a web page) in your browser and you are not authorised on the the network, you will only ever see the page which says "Welcome to Free -wifi network. here are our terms etc..."
I'm not interested in abusing the service for free wifi - I want to know what information (HTTP transactions) is exposed/insecure when I use this service and what persistent changes they are making to my system, i.e. hiding cookies in my browser (by using anonymous hosts).
So I connected to their wifi access point, at which point they assign my computer an IP address and DNS service etc.
$ nmcli device show wlp2s0 IP4.ADDRESS: 192.168.12.199/22 IP4.GATEWAY: 192.168.12.1 IP4.ROUTE: dst = 169.254.0.0/16, nh = 0.0.0.0, mt = 1000 IP4.DNS: 123.123.xxx.xxx IP4.DNS: 123.123.xxx.xxx
I then attempt a HTTP request - try to "browse to a webpage" while not yet authorised. (Being HTTP this request/response will be unencrypted, no identity check on the server handling the request [via. SSL certificate authority confirmation via public / private key signatures etc])
$ curl 'http://placeimg.com/' -H 'Host: placeimg.com' -H 'User-Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:47.0) Gecko/20100101 Firefox/47.0' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8' -v -L * Trying 220.127.116.11... * Connected to placeimg.com (18.104.22.168) port 80 (#0) > GET / HTTP/1.1 > Host: placeimg.com > < HTTP/1.1 302 Found < Location: http://22.214.171.124:80/reg.php?ah_goal=auth-tc.html&ah_login=true&url=E2B8F357812412D8 < * Ignoring the response-body * Connection #0 to host placeimg.com left intact * Issue another request to this URL: 'http://126.96.36.199:80/reg.php?ah_goal=auth-tc.html&ah_login=true&url=E2B8F35792412D8' * Connection 0 seems to be dead! * Closing connection 0 * Trying 188.8.131.52... * Connected to 184.108.40.206 (220.127.116.11) port 80 (#1) > GET /reg.php?ah_goal=auth-tc.html&ah_login=true&url=E2B8F8 HTTP/1.1 > Host: 18.104.22.168 > User-Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:47.0) Gecko/20100101 Firefox/47.0 > Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8 > Accept-Language: en-US,en;q=0.5 > < HTTP/1.1 200 OK < Date: Mon, 08 Aug 2016 02:50:16 GMT < Connection: keep-alive < Transfer-Encoding: chunked < Content-type: text/html; charset=utf-8 < <!DOCTYPE html> <html xmlns="http://www.w3.org/1999/xhtml"> <head> <meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /> <meta name="viewport" content="width=device-width"/> <title>Login</title> <h1>WiFi - Free Internet Access</h1> This service is provided primarily for the purposes of accessing Email and light web browsing .. .... </div> </body> </html>
I've truncated and redacted the output, but essentially its seems, when my browser (curl) makes a request the network is providing access to a real DNS server, I tested that further by running this -while connected to the service.
$dig +short google.com 22.214.171.124 126.96.36.199 ... (basically real ip addresses for google.com)
So its seems the service is providing real, truthful information about DNS,
But even though the IP address my browser is addressing seems valid, the HTTP response coming back, is not from the intended server (its a server that has been inserted by the WiFi service). The response is
302telling my browser make a new request, to the
188.8.131.52/reg.phpserver and URL
- the response from that server/URL is the "captive portal" page.
When make a HTTPS request
$ curl 'https://google.com/' -H 'Host: google.com' -H 'User-Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:47.0) Gecko/20100101 Firefox/47.0' -H 'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8' -H 'Accept-Language: en-US,en;q=0.5' -v -L * Trying 184.108.40.206... * connect to 220.127.116.11 port 443 failed: Connection refused * Trying 2404:6800:4006:800::200e... * Immediate connect fail for 2404:6800:4006:800::200e: Network is unreachable * Trying 2404:6800:4006:800::200e... * Immediate connect fail for 2404:6800:4006:800::200e: Network is unreachable * Failed to connect to google.com port 443: Connection refused * Closing connection 0 curl: (7) Failed to connect to google.com port 443: Connection refused
The request outright fails - I'm assuming because my computer won't accept a connection from a remote host it can't authenticate?
So my question is, is this a "man in the middle" type interception at the TCP/IP level? in other words, when the network stack on my computer tries to establish a TCP connection with the
placeimg.com server, by starting a 3 way handshake, is the Wifi services gateway or network node responding to my SYN with an ACK/SYN and saying yes I'm that server you wanted to talk to - lets complete our TCP connection, then once established its pre-programmed to automatically send the HTTP redirect and then, because my browser is actually generating the next request to
18.104.22.168 itself - all the "funny business" can stop and its all normal HTTP/HTML behaviour from that point forward. (ie. standard form filling and submissions as you would on any normal website)
If not, how is this gateway intercepting and re-routing my requests while I am an "unauthenticated" user?