TELCOM 2310 Fall 2024
Project 2: Internet Measurement and Analysis
In this project you will become more familiar with some of the basic tools used for Internet
measurements. In particular, you will use traceroute, which, as we saw in lab 3, is used to identify
the end-to-end path followed between two machines (or more precisely, between two interfaces).
The output of traceroute is essentially a list of the routers traversed on the path from the source to
the destination. It also provides information about packet loss and the delay experienced at each
hop in the route1.
This project consists of the three following parts:
1. Measurements/Data collection: You will use traceroute to collect measurement data.
Specifically, you must pick 5 IP destinations (e.g., www.google.com, www.pitt.edu, e tc.) and
use traceroute to identify the path followed to these destinations. You should collect all of
your measurements using the same source machine (i.e. the machine that you run traceroute
from); this can be your home computer, a specific lab computer you have access to at Pitt, a
specific public traceroute server, a particular GENI node, etc.).
For each of your 5 destinations, you will collect multiple measurements spread across time.
For example, you can choose to make x measurements per destination per day for y days. This
will provide you with x ∗ y datapoints. However, be careful to ensure that these datapoints
are spread across a sufficient time pe riod: you ma y no t co llect al l x ∗ y da tapoints at once
(e.g. within the same hour). Make sure you start the project early enough to give
yourself enough time to collect your data.
Notes on choosing a source machine: If you use your home computer/laptop for this purpose,
I would recommend using a wired connection when performing the traceroutes rather than
doing them over Wifi; the routes you observe should be the same in e ither case, but latency
measurements are likely to be more variable and harder to interpret for wireless connections.
Certain ISPs may also block traceroute traffic (if you run traceroute and see the source and
destination, but no intervening routers, this is likely the case).
Notes on choosing destinations: If you use a domain name as your traceroute destination (e.g.
www.google.com), note that it may not always correspond to the same IP address. Recall
our discussion of CDNs, and how websites may be served from replicated webservers in many
different geographic locations, with different IP addresses. This is perfectly fine, but you
1For more details on traceroute you can see: http://linux.die.net/man/8/traceroute and https:
//www.lifewire.com/traceroute-linux-command-4092586 (for Linux users) and https://www.lifewire.com/
tracert-command-2618101 (for the Windows implementation). Linux instructions will apply to most Mac users
as well; talk to the instructor if you have any problems.
1
should be aware of it in your analysis to distinguish the case where the route being taken to
some IP address has changed vs the case where the destination IP address has changed.
2. Data analysis: As we have discussed in class, the statistical multiplexing approach of the
Internet means that packets traveling between two specific points can follow different paths,
even if they belong to the same UDP flow or TCP connection. The goal of this project is
to use the data you collect in Step 1 of the project to analyze the stability of these paths at
different granularities:
(a) The output of traceroute is a path at the router-level. The first part of the analysis phase
of the project requires you to translate each router-level path to an autonomous-system-
level path2. You can use the whois database to translate the router IP addresses from
your traceroute output to AS numbers: https://team-cymru.com/community-services/
ip-asn-mapping/. See the website for more details (and instructions on querying for
multiple IP addresses simultaneously), but a simple example for a single query (for IP
address 68.86.100.25) might look like:
whois -h whois.cymru.com " -v 68.86.100.25"
with output:
AS | IP | BGP Prefix | CC | Registry | Allocated | AS Name
7922 | 68.86.100.25 | 68.80.0.0/13 | US | arin | 2002-01-28 | COMCAST-7922, US
Here, the AS number for this IP address is 7922, and we can see that this AS is managed
by Comcast (AS Name = COMCAST-7922).
(b) After translating router-level paths to AS-level paths, you must analyze the stability of
the paths both at the router-level and the AS-level. Sample questions to address are:
“What is the most dominant path?” “How often do paths change?” “Which part of the
path changes most?”, etc.
(c) Finally, you must analyze the stability of the delay for both the end-to-end paths (from
your source machine to each of the 5 destinations you picked) and for the individual hops
you identify through traceroute (at the AS-level as well as the router-level). Of course,
to be able to analyze individual hops, you need to have the same hop appear multiple
times in your dataset, so this may not be possible for every hop.
Notice that we are not specifying the exact metrics you should use to capture the stability of
the paths you observe or the stability of the delay. It is part of your assignment to identify
metrics that provide insight into the stability. You also will need to choose the granularity of
your measurements for the data collection phase.
3. Report: You must submit a clearly written report that contains the following:
(a) The goal of this project (introduction)
(b) The methodology for your measurements (report on step 1)
(c) The metrics you picked for your data analysis, the reason you picked these metrics, and
the actual results (report on step 2)
2Recall that, at a high level, an autonomous system (AS) is a set of routers that belong under the same adminis-
trative authority
2
(d) Your conclusions about path and delay stability based on your analysis results
The project will be graded as follows:
• 30% data measurements
• 40% data analysis (metrics picked and justification, data analysis and visualization/presentation).
Note that this must include some analysis of (1) router-level path stability, (2) AS-level path
stability, and (3) delay stability.
• 30% report (clear presentation, structure, conclusions)
Deadline: Wednesday, Dec. 09, 2024, 5:59 PM EDT. No extensions are possible since it will
be end of the term, so please start early!