TELE 3118: Network Technologies: Lab 4 [5 points]

The objective of this lab is for you to get some familiarity with routing in the Internet. This requires you to log into real Internet routers, and observe the routing tables and AS (Autonomous System) paths. Additionally, you will use the traceroute facility to deduce what routes packets are taking in the Internet. You will find links to useful resources such as route-servers and looking-glasses at the traceroute website.

Route-servers are Internet routers you can log in to and inspect the various tables. For example, log into the AT&T route-server by typing "telnet route-server.ip.att.net". You will get a command-line interface (CLI); at any point you can type a "?" to get a listing of possible completions to the commands you are typing. For example, "show ip ?" will give you a listing of the various IP-related attributes that can be inspected (this includes interface tables, ARP tables, routing tables, etc). Take the time to go through the options and familiarise yourself with the CLI. You can also experiment with numerous route-servers listed at the traceroute website above.

Looking-glasses let you perform trace-route operations from remote sites. For example, you can go to the SAVVIS looking-glass and try a traceroute to a host in UNSW, e.g. 149.171.92.217. Note the path, as well as the round-trip-time to each hop on the path - can you guess which links are trans-oceanic?

What to submit:

You submission should clearly answer the following questions (note that most questions have multiple parts):
  1. [0.5] Pick a route-server overseas (e.g. AT&T or Savvis), stating your choice clearly in your answer. Now log in to the route-server, and use the CLI to view the interface IP address table and ARP table. What are the IP addresses/masks of each interface the route-server has?
  2. [0.5] Now examine the routing table. How many routes are in the routing table (remember to count networks and subnets)? - this gives you an idea of the Internet routing table size. How many of these routes are static, from an IGP, and from BGP respectively?
  3. [0.5+0.5] Using the BGP routing tables at the route-server, determine the AS-path (i.e. list of AS numbers) that a packet destined to a UNSW destination (e.g. 149.171.92.217) would take. For this AS-path, now identify by name the Autonomous Systems on the path (in other words, name the ISPs a packet from the route-server to UNSW would have to go through). [Hint: the CLI command "show ip bgp x.x.x.x" performs a longest-prefix match on the destination IP address x.x.x.x in the BGP routing table and shows you all AS paths, including the best one. To determine the AS name given the AS number, you can use the "whois" command on Linux, or the web-based list of AS names.
  4. [0.5+0.5] Now using the CLI on the route-server, perform a traceroute to the UNSW address you chose above. For each router on the path, determine the AS to which it belongs (the whois tool can be used to determine the AS to which an IP address belongs). Verify whether this matches the AS-path from the previous part or not.
  5. [0.5+0.5+0.5] The traceroute web-site lists around 16 traceroute gateways in Australia. Of these, find the gateway in Australia from which the delay to (any host/domain in) UNSW is lower than 20 msec (if no such gateway exists, find the lowest), and find a gateway from which the delay to UNSW is over 200msec (if no such gateway exists, find the highest). By noting the paths, can you explain why there is such a large variation in the delays?
  6. [0.5] Based on the previous part, and by possibly doing more trace-routes and whois lookups to hosts in UNSW, deduce who UNSW's primary ISP is. Does UNSW connect to more than one ISP?
Directions: The answers to the above need to be submitted on this sheet to the lab instructor at the end of your lab session.