Welcome to the homepage of Advanded Networking (ANET), edition
2023/2024. ANET is a course for M.Sc. students at the University of
Twente that enables you to understand, discuss, and evaluate advanced
internetworking concepts, such as secure inter-domain routing,
multi-path communication, path control, and experimental Internet
architectures. ANET is a first quartile course (September through
November).
The main audience of this site are students who enrolled for
the 2023/2024 edition of ANET. However, it might also be of
interest to potential future students and perhaps to fellow teachers
around the globe who’d like to reuse parts of our format and/or content
at their institution (or who’d like to provide feedback :-)
Students: please read this page carefully, so you
know what we expect from you, and what you can expect from us. We’ll
provide a summary at the ANET introduction lecture on Wed Sep 6, 2023
(see the schedule below or check https://rooster.utwente.nl/).
The section Additional Reading points to papers and reports on advanced
networking concepts that we won’t discuss in class, but that we think
are nontheless very interesting. Feel free to check them out and discuss
them with us if you want.
Overview
ECTS
5 (140 hours)
Code
201700077-1A
Prerequisites
Introductory course in computer networking, such as the bachelor
module Network Systems at the UT
Teachering team
dr.ir. Pieter-Tjerk de Boer (University of Twente) prof.dr.ir.
Geert Heijenk (University of Twente) prof.dr.ir. Roland van
Rijswijk-Deij (University of Twente and NLnet Labs) prof.dr. Cristian
Hesselman (SIDN Labs and University of Twente)
Ranked top five of papers you’d like to blog about (1st, 2nd, etc.):
Wed Sep 6, 2023
Individual and group test: at each lecture
Blog: one week after the lecture in which you
presented the paper
Lab assignment: by the end of the last lab session
(see ANET schedule)
Notification of grades: two weeks after the last
lecture, so around Nov 22, through Canvas
Enrollment
You can only sign up for ANET through
OSIRIS,
we do not accept registrations via email or other channels. We urge you
to scan ANET’s list of papers before you sign up, so
you know what to expect. Should you end up unsubscribing from ANET, then
please unregister through OSIRIS. In this case, we appreciate it if you
share your feedback as to why you unregistered with the ANET coordinator
(email above) so we can use it to further improve the course.
ANET is a highly interactive course in which you
study research papers on advanced networking concepts, engage in
technical discussions with your peers (“group-based learning”), and
present a paper you studied in depth. The number of seats is
limited to 16 because we can only fit two presentations
in one lecture.
We admit people on a first-come-first-served basis using the OSIRIS
timestamp of your registration. We’ll put you on a waiting list if more
than 16 students sign up and let you know if a seat becomes available.
Course updates
We’ll keep you up to date through this website when it comes
to ANET’s content and Canvas for announcements and
uploading/archiving deliverables.
A separate public website is important to us because we’d like to
share the course’s format with other universities and students so they
can potentially learn from our work (open education). Also, at SIDN Labs
we’re proud of our contribution to courses like ANET and we’d like to
use this site to underscore the importance of companies helping to
educate the next generation of engineers and researchers. Finally, the
UT is a public institution, which we believe means it should share its
output with the Dutch and European society and beyond as much as
possible.
Prerequisites
You need to have passed an introductory course in computer networks,
such as the bachelor module Network Systems at the University of Twente.
It is your responsibilty to assess if you have the
required background knowledge for the course (see Prequisites). This is
also why we highly recommend scanning the ANET papers before you enroll
(see Enrollment). Contact the course coordinator if you have any
questions.
Lectures
ANET offers nine interactive lectures, two guest lectures, and two
lab sessions. We'll start with an introduction to ANET on Wed
Sep 6 (see Schedule) to inform you of what will be expected
from you and to answer any questions you may have.
The references (between square brackets “[..]”) point to the papers
at the bottom of this page.
Regular lectures
Six of the eight interactive technical lectures focus on advanced
networking concepts such as inter-domain routing [BGP] and multi-path
TCP [MTCP1] [MTCP2], while the remaining two provide a more high-level
perspective and concentrate on the design of the Internet and potential
alternative architectures, such as SCION [SCION] and RINA [RINA]. Our
motivation for including this last kind of topic is to
enable you to understand the Internet’s design goals and broaden your
view beyond the TCP/IP architecture.
Each lecture revolves around a specific theme, which
together give you a broad overview of the core functions of inter-domain
networking [Shosh]. We picked the topics “going up the stack”: from
programmable networks (the hardware), to BGP security (routing and
addressing), DNS security and privacy (naming), transport-level
protocols (QUIC and multi-path communication), a specific networking
application (data centers), and Internet architectures (a broad view on
inter-domain neworking). The other reason for selecting these topics is
that your teachers are conducting research in these areas, which means
they can tell you everything about it (and of course we love to talk
about the work we do :-)
Each lecture covers three papers. An
introductory paper to give you a basic understanding of
a particular topic, followed by two advanced papers
that explore the topic in more depth. We test your understanding of the
introductory papers through a closed book multiple-choice
test that you will need to make in class. You’ll first take the
test individually, and then the same test again in a
group with 2-3 of your fellow students. The purpose of the
group test is to enable you to learn from your peers by discussing the
questions of the test with them.
We test your understanding of one of the course’s advanced
papers by asking you to write a review about it in the form of
a blog and present it in class. We’ll
publish the best blog on the ANET website, of course only with the
author’s explicit consent. We won’t publish the other blogs. Please
check the Assessment section and the section deliverable #2 (blog) for
more details.
The papers roughly cover well over a decade of networking
research, with one paper from the Internet’s proverbial “stone
age” (1995). This is by design because we aim to help you understand
generic network architectures and principles (e.g.,
software defined networking), which have a longer lifetime than today’s
latest and greatest topics.
You must attend all lectures because that’s
when the tests will take place and because of ANET’s interactive nature
(more on that in the section Detailed timetable for regular
lectures).
Guest lectures and lab
sessions
In addition to the regular lectures, we also offer two guest
lectures (see section Guest Lectures), an P4
introduction, and two P4 lab sessions. The
goal of the guest lectures is to give you a flavor of operational
network infrastructure (current and experimental) and of ongoing
research in one of the areas we address in ANET (e.g., routing).
Schedule
Table 1 shows the schedule of ANET. The regular lectures (Rx) usually
take place on Wednesdays, the P4 lab sessions (Lx) and guest lectures
(Gx) on Mondays. Note that there may be exceptions (e.g., depending on
the availability of guest lecturers and your teaching team), so make
sure to regularly check the Time Tables site regularly: https://rooster.utwente.nl/.
The all-afternoon lab sessions enable you to allocate two large
chuncks of time to the P4 excersises, but we strongly advice you
also work on the P4 work at home and not just at the
lab sessions (see Deliverable #4: P4 lab assignment).
Internet architectures I [SCION] [NDN] [FABRID] Teacher: Cristian
Hesselman
L2
Mon Oct 23 13:45-17:30
VR 559
P4 lab session #2 Teacher: Nathan Djojomoenawie
R8
Thu Oct 26 15:45-17:30
HB 2E
Data Center Networking [DCN1] [DCN2] [B4] Teacher: Geert Heijenk
R9
Wed Nov 1 13:45-15:30
VR 212
Internet architectures II [DARPA] [TROTS] [TUSSLE] Teacher:
Cristian Hesselman
Detailed timetable for
regular lectures
Table 2 provides a detailed schedule for the regular lectures. We
apologize for the micromanagement, but this is important to make the
best of our time.
Table 2. Detailed timetable per on-campus lecture.
Time
What
10:45
Arrival, put your cell phone in your bag, pick up hardcopy of tests at
teacher’s desk, sit down
10:45-10:55
Individual test of introductory paper (closed book) Your teacher
will pick up the tests when everyone is done
10:55-11:00
Organize into groups (your teacher divides you across groups)
11:00-11:20
Group test of introductory paper (closed book), including open
question Your teacher will pick up the tests when everyone is done
A total of 8 multiple-choice tests on introductory
papers for a particular topic (e.g., multi-path networking), to be
completed individually and in groups in class
A blog of at most 1.500 words in which you review one
of the advanced papers. We’ll publish the best one on the ANET site,
with the author’s consent. The others we won’t publish.
A presentation of 15 minutes about that paper at one of
the lectures, including 5 minutes of discussion
A P4 program that configures the packet handling
functions of a P4-programmable network switch, to be carried out
individually at one of the two lab sessions or at home.
We expect you to both actively participate in the discussions for the
group tests, in the plenary discussion, and in the Q&A after each
talk. Attendance of the lectures is therefore
mandatory.
Assessment
We asses to what extent you attained ANET’s learning outcomes (see
Background) based on academic papers and RFCs you will need to study and
a lab assignment you will need to carry out.
To pass ANET, your overall score will need to be 5.5
or higher, which we calculate as follows:
((average score of your 8 individual tests)*25% + (average score
of your 8 group tests)*25% + (score of your blog)*40% + (score of your
presentation)*10%) * (score of your lab assignment)
The scores of the tests, blog, and presentation are between 1 (worst)
and 10 (best). The score of the lab assignment is either 1 (pass) or 0
(fail).
There are more details on how we grade each of your deliverables
(group tests, blog, presentation, and lab assignment) in the sections
below.
Deliverable #1: eight
multiple choice tests
Each ANET lecture involves a multiple-choice test to
assess your understanding of the introductory paper on the topic that we
discuss in a particular lecture (see the section Schedule). The test is
closed book and you will need to make it in
class, first individually and then the same
test in a group with 2-3 of your fellow students. The
goal of the group-based test is to enable you to learn from your peers
and further increase your understanding of the paper. The group-based
test also has one open question in which we ask you to
formulate the main takeway of the paper in one sentence of at most 25
words.
After the tests, we’ll take 20 minutes to discuss the paper the whole
group to enable everyone to learn from each other.
We calculate your grade for a particular test as follows:
Grade of individual part of the test = maximum of
((S-G)/(Q-G))*9+1 and 1
Where S is the number of answers you got right, Q is the number of
questions in the test, and G is the “guessing factor”. The latter is the
number of answers you could have guessed correctly if you hadn’t read
the paper, which is usually Q/4 (most tests have 4 answers you can
choose from).
The open question of the group-based tests will get you one point,
which means that we grade those tests as follows:
Grade of group-based part of the test = maximum of
((S-G)/(Q-G))*8+O+1 and 1
Where O is 1 if you got the open question right and 0 if you
didn’t.
Deliverable #2: blog
Your second deliverable is a blog in which you
review one of the advanced papers in your own
words in at most 1.500 words. The goal of your blog is to
enable readers to assess if they’d like to read the full paper based on
your review.
The topics you discuss in the blog should capture the core of what
the paper is about. Table 3 lists a few of the topics you can discuss in
your blog.
Table 3. Example topics to blog about.
Design paper (e.g., [SCION])
Measurement paper (e.g., [DNS-SP])
What is the problem that the authors aim to solve?
What is the problem that the authors aim to solve?
What requirements do the authors articulate for their work?
What methodology and experimental setup do the authors use?
What does the high-level design and operation of their proposed system
look like?
What are their key findings and conclusions?
How does the design address the requirements?
How do they propose others use their measurement study?
What are the pros and cons of the authors’ work and why?
What are the pros and cons of the authors’ work and why?
What would you do differently?
What would you do differently?
Would you recommend the paper to interested readers?
Would you recommend the paper to interested readers?
Make sure your blog is self-contained, which means
that folks who haven’t read the paper should be able to understand it.
Your target audience are readers with a background in computer
networking, so you don’t need to explain basic networking
concepts.
Your blog must begin with a section in which you
explain the paper’s three main takeways in at most 150
words. The “Key Insights” box on page 1 of paper [SCION] has an example
that might provide inspiration.
At the end of your blog, you’ll need to
briefly reflect on the process you followed to write
the blog in at most 100 words. For example, hwo did you go about
studying the paper and understanding its contents, and how did you write
the blog. Please also indicate how you incorporated the feedback you
received at your presentation. The reflection does not add to the word
count of your blog.
You also need to include an appendix that says:
“I wrote this blog myself. I used [TOOL/SERVICE] exclusively to
improve the language of the blog and not to generate content that I had
not already written myself. I fully reviewed the [TOOL/SERVICE]-revised
versions of the blog and take full responsibility for its
content.”
OR
“I did not use any artificial intelligence tools to write my
blog.”
Another source of inspiration may be prof. Aiko Pras’ guide
on how to write a scientific paper. That’s different from a blog, but
the video might help you structuring your blog anyway.
Referencing, citing, and
quoting
If you want to literally include text from the paper (or other
sources) in your blog, then you need to reference and
quote it. This includes sources where you are an author. During
the first lecture, we will briefly discuss the differences between
citing, quoting, and copying (slide of 19 of this
presentation).
If you want, you can use references in your blog rather than
hyperlinks to refer to external sources. We’ll exclude them from the
wordcount, and we also won’t count the blog’s title and figure and table
headings.
You can reuse graphs from the paper if you want, but make sure
clearly indicate that they’re from the paper and not your own.
Note that Canvas will automatically check your blog for plagiarism
when you submit it. As per the university’s policy, no forms of
plagiarism are tolerated.
Please contact us if you have questions regarding citing and
quoting.
Who writes about which paper?
Table 4 shows who blogs about which paper.
Table 4. Bloggers and papers.
First name
Blogs about
Lynn
[TUSSLE]
Ferry
[MTCP2]
Félix
[TROTS]
Max
[BGP2]
Tim
[DoH]
Pedro
[NDN]
Alimzhan
[QUIC2]
Publication of the best blog
We’ll publish the blog with the highest grade on the ANET site after
the course, so other students you can learn from your work (and you can
show it off to your friends :-)
Of course, we’ll only publish the blog with the author’s explicit
consent.
Evaluation criteria
We will evaluate your blog based on the following
criteria:
Understanding: how well did you understand the paper, for instance in
terms of the problem it aims to solve and the paper’s key points?
Analysis: to what extent did you provide a critical analysis of the
paper, for instance in terms of the strengths/weaknesses of the work,
limitations of the proposed solution/approach, and potential
improvements?
Clarity: structure, language, and readability of the blog.
Grading process
The ANET teacher who gives a particular lecture will evaluate the
blogs about the papers of that lecture. In addition, one of the other
teachers will review your blog for a cross-check. They both use the
evaluation criteria listed above to grade your blog.
We’ll be evaluating all of your blogs in one go at the end of the
course. This means that you’ll get your grade around two weeks after the
last presentation. That’s because that person has 1 week to submit their
blog and we’ll need a few days to evaluate your work.
The reason for reviewing and evaluating all blogs in one go is that
this approach will allow for a fairer assessment because we have the
overview of all blogs, plus that it’s more efficient on our part.
Deliverable #3: presentation
You’ll need to present the paper that you wrote a blog about to your
peers in at most 15 minutes, consisting of 10 minutes
of speaking time and 5 minutes of questions and discussion.
Your presentation must start with your three main
takeaways of the paper.
Your lecturers will score your presentation based on clarity,
structure, and how well you responded to questions. Your fellow
students will do the same through a feedback form that we’ll hand out at
the beginning of each lecture. Your lecturers will use this feedback to
round off scores. You must return the feedback forms before leaving the
room.
Here’s a few How To’s for giving a talk:
Ian Parberry: How to Present a Paper in Theoretical Computer Science - A
Speaker's Guide for Students.
Simon L. Peyton Jones, John Hughes, John Launchbury: How to give a good
research talk.
Deliverable #4: P4 lab
assignment
The lab assignment involves you programming the packet handling
functions of a (simulated) open programmable router using the
domain-specific language P4 [P4a], which we believe is an important
enabler for an open programmable Internet infrastructure [RI].
We will discuss P4 in one of the first lectures and we’ll provide
detailed information on the lab assignment through an introductory
lecture (see Schedule). The Teaching Assistant is available for
questions.
You will need to carry out the P4 assignment
individually at one of the two lab sessions or at home.
When you’re done, you need to show your results to the Teaching
Assistant in one of two lab sessions (see Schedule).
The teaching assistant will sign off your P4 assignments
if:
Your P4 code runs and shows the expected behavior, and
You’re able to explain the Teaching Assistant what’s going on and why,
and
You added comments to your P4 code explaining why you programmed the
router the way you did
We strongly advice that you also work on the P4 lab
assignment at home and not only at the lab sessions
(see Schedule). That’s because it might take some of you more than the
two lab sessions to finish the work. Also keep in mind that the Teaching
Assistant needs to help multiple students at the lab sessions, so the
person might not always be immediately available for you.
A few P4 tutorials that you might find useful are:
Please submit your blog, presentation, and P4 code through Canvas
(unzipped) so we can confirm that you yourself carried out the work
(SimCheck).
You do not need to submit the multiple-choice tests through Canvas
because we’ll get them from you in class.
Resits
In some cases, you can take a resit for ANET’s deliverables, such as
when you missed a test because you were ill or if you failed a
deliverable. Please check with the course coordinator if you think this
applies to you.
We offer resits as follows:
Multiple-choice test: we’ll enable you to take an oral exam instead,
with the examiners being the teacher of the lecture and one other member
of the teaching team.
Blog/presentation: we’ll enable you to write a new blog about another
paper and present it to a few members of the teaching team.
Guest lectures
We provide two guest lectures by industry experts to further broaden
your horizon on inter-domain networking. The lectures will be open for
everyone (ANET students, other students, and university staff).
Lecture
#1: a sample of current routing vulnerabilities and how we may hack to
live with (some of) them
Lecturer: prof. Cristel Pelsser, UCLouvain, Belgium
Abstract: BGP, the routing protocol that runs between large networks
was not designed with security in mind. In this class, I’ll present some
of the attacks that are possible ranging from hijacks, blackholing using
communities, path manipulation, and BGP session resets. I’ll then move
on to present some of the proposal aimed at detecting attacks on the
protocol. The difficulty in detecting issues in inter-domain routing
lies in that BGP hides information on the details of the Internet
topology. This property is a given as it is at the root of the
scalability of the protocol. How do we aim to fill the holes to improve
detection and better understand the Internet?
Lecture #2: submarine
networking
Lecturer: Dr. Roeland Nuijts, Ciena Networks
Submarine Cable Industry Trends - lecture on the growing importance
of transoceanic intercontinental internet transport capacity, and how to
overcome the technical and geopolitical challenges to meet growth
expectations. Topics will include but not be limited to: the Shannon
limit, new (multi-core) fiber types, SDM, Sustainability issues,
Regulatory challenges and Cabotage (Cable Sabotage).
Canvas conferences
ANET lectures take place on campus, but we might use Canvas
conferences for ANET guest lectures.
You can join the online ANET sessions by going to the ANET course
page on Canvas, browse to Conferences, and join the session. We aim to
open sessions at least 10 minutes prior to its start. Please join with
your microphone muted. Do not join in listen-only mode.
Survey
At the end of the course, the university’s Quality Assurance
department will send each of you an online survey to ask how you
experienced ANET and how you think we could further improve it. We
kindly request a few minutes of your time to fill it out so we can
learn from your feedback and further improve ANET next
year. Your fellow students who will take ANET next year will thank you
for it!
At the beginning of the course, we’ll inform you what changes we made
based on the feedback we received from last year’s students. (We’ll
inform them those changes as well, so they know what we did with their
feedback.)
ANET papers and RFCs
We use a total of 24 papers and Request for Comments
(RFCs) for ANET. Please check the Lectures section on why we
picked these papers and how they interrelate.
Programmable Networks
Introductory paper:
[SDN] B. Nunes Astuto, M. Mendonca, X. Nam Nguyen, K. Obraczka, T.
Turletti, “A Survey of Software-Defined Networking: Past, Present, and
Future of Programmable Networks”, IEEE Communications Surveys &
Tutorials, Vol. 16, Issue 3, Third Quarter 2014, https://www.comp.nus.edu.sg/~chanmc/cs6204/week1-survey.pdf
Advanced papers:
[P4a] P. Bossharty, D. Daly, G. Gibby, M. Izzardy, N. McKeownz, J.
Rexford, C. Schlesinger, D. Talaycoy, A. Vahdat, G. Varghesex, and D.
Walker, “P4: Programming Protocol-Independent Packet Processors”, ACM
SIGCOMM Computer Communication Review, Volume 44, Issue 3, July 2014,
pp. 87-95, https://www.sigcomm.org/sites/default/files/ccr/papers/2014/July/0000000-0000004.pdf
[P4b] S. Knossen, J. Hill, P. Grosso, “Hop Recording and Forwarding
State Logging: Two Implementations for Path Tracking in P4”, 2019
IEEE/ACM Innovating the Network for Data-Intensive Science (INDIS),
November 2019, https://ieeexplore.ieee.org/document/8940312
BGP Security
Introductory paper:
[BGP1] S. Goldberg, “Why is it taking so long to secure internet
routing?”, Communications of the ACM, Vol. 57, Issue 10, October 2014,
pp 56–63, https://doi.org/10.1145/2659899
Advanced papers:
[BGP2] T. Chung, E. Aben, T. Bruijnzeels, B. Chandrasekaran, D.
Choffnes, D. Levin, B. M. Maggs, A. Mislove, R. van Rijswijk-Deij, J. P
Rula, N. Sullivan, “RPKI is Coming of Age: A Longitudinal Study of RPKI
Deployment and Invalid Route Origins”, Proceedings of the Internet
Measurement Conference, October 2019, pp. 406–419, https://doi.org/10.1145/3355369.3355596
[BGP3] J. Kristoff, R. Bush, C. Kanich, G. Michaelson, A. Phokeer, T. C.
Schmidt, M. Wählisch, “On Measuring RPKI Relying Parties”, Proceedings
of the ACM Internet Measurement Conference, October 2020, pp. 484–491,
https://doi.org/10.1145/3419394.3423622
[DNS-SP] L. Zhu, Z. Hu, J. Heidemann, D. Wessels, A. Mankin, N. Somaiya,
“Connection-oriented DNS to Improve Privacy and Security”, Proceedings
of the 2015 IEEE Symposium on Security and Privacy, 2015, https://nymity.ch/tor-dns/bibliography/pdf/Zhu2015a.pdf
Multi-path communication
Introductory paper:
[MTCP1] C. Paasch, O. Bonaventure, “Multipath TCP”, Communications of
the ACM, Vol. 57, No. 4, April 2014, pp 51–57, https://doi.org/10.1145/2578901
Advanced papers:
[MTCP2] D. Wischik, C. Raiciu, A. Greenhalgh, M. Handley, “Design,
implementation and evaluation of congestion control for multipath TCP”,
Proceedings of the 8th USENIX conference on Networked systems design and
implementation, 2011, https://www.usenix.org/legacy/event/nsdi11/tech/full_papers/Wischik.pdf
[MTCP3] Z. Morley Mao, S. Sen, “An in-depth understanding of multipath
TCP on mobile devices: measurement and system design”, Proceedings of
the 22nd Annual International Conference on Mobile Computing and
Networking (MobiCom ’16), New York City, New York, Oct 2016, https://dl.acm.org/citation.cfm?id=2973769
QUIC
Introductory paper:
[QUIC1] A. Langley, A. Riddoch, A. Wilk, A. Vicente, C. Krasic, D.
Zhang, F. Yang, F. Kouranov, I. Swett, J. Iyengar, J. Bailey, J.
Dorfman, J. Roskind, J. Kulik, P. Westin, R. Tenneti, R. Shade, R.
Hamilton, V. Vasiliev, W.-T. Chang, Z. Shi, “The QUIC Transport
Protocol: Design and Internet-Scale Deployment”, SIGCOMM 2017, August
2017, https://dl.acm.org/doi/pdf/10.1145/3098822.3098842
Advanced papers:
[QUIC2] Arash Molavi Kakhki, Samuel Jero, David Choffnes, Cristina
Nita-Rotaru, Alan Mislove, “Taking a Long Look at QUIC - An Approach for
Rigorous Evaluation of Rapidly Evolving Transport Protocols”, Internet
Measurement Conference (IMC2017), November 2017, https://dl.acm.org/doi/pdf/10.1145/3131365.3131368
[QUIC3] T. Viernickel, A. Froemmgen, A. Rizk, B. Koldehofe, R.
Steinmetz, “Multipath QUIC: A Deployable Multipath Transport Protocol”,
2018 IEEE International Conference on Communications (ICC), May 2018, https://ieeexplore.ieee.org/abstract/document/8422951
Data Center Networking
Introductory paper:
[DCN1] T. Wang, Z. Su, Y. Xia and M. Hamdi, “Rethinking the Data Center
Networking: Architecture, Network Protocols, and Resource Sharing”, IEEE
Access, vol. 2, pp. 1481-1496, December 2014, https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=6990724,
up to (not including) Section V.A
Advanced papers:
[DCN2] Leon Poutievski, Omid Mashayekhi, Joon Ong, Arjun Singh, Mukarram
Tariq, Rui Wang, Jianan Zhang, Virginia Beauregard, Patrick Conner,
Steve Gribble, Rishi Kapoor, Stephen Kratzer, Nanfang Li, Hong Liu,
Karthik Nagaraj, Jason Ornstein, Samir Sawhney, Ryohei Urata, Lorenzo
Vicisano, Kevin Yasumura, Shidong Zhang, Junlan Zhou, and Amin Vahdat,
“Jupiter evolving: transforming google’s datacenter network via optical
circuit switches and software-defined networking”, ACM SIGCOMM 2022,
August 2022, https://dl.acm.org/doi/abs/10.1145/3544216.3544265
[B4] Sushant Jain, Alok Kumar, Subhasree Mandal, Joon Ong, Leon
Poutievski, Arjun Singh, Subbaiah Venkata, Jim Wanderer, Junlan Zhou,
Min Zhu, Jonathan Zolla, Urs Hölzle, Stephen Stuart and Amin Vahda, “B4:
Experience with a Globally-Deployed Software Defined WAN”, SIGCOMM 2013,
August 2013, https://dl.acm.org/doi/pdf/10.1145/2486001.2486019
Internet architectures, part
I
Introductory paper:
[SCION] D. Barrera, L. Chuat, A. Perrig, R. M. Reischuk, and P.
Szalachowski, “The SCION Internet Architecture”, Communications of the
ACM, Vol. 60, No. 6, June 2017, https://scion-architecture.net/pdf/2017-SCION-CACM.pdf
[TROTS] J. McCauley, Y. Harchol, A. Panda, B. Raghavan, and S. Shenker,
“Enabling a permanent revolution in internet architecture”, Conference
of the ACM Special Interest Group on Data Communication, August 2019,
pp. 1-14, https://dl.acm.org/doi/pdf/10.1145/3341302.3342075
[TUSSLE] D.D. Clark, J. Wroclawski, K.R. Sollins, and R. Braden, “Tussle
in cyberspace: defining tomorrow’s Internet”, IEEE/ACM Transactions on
Networking, Vol. 13, Issue 3, June 2005, https://ieeexplore.ieee.org/document/1458757
Additional reading
[ICING] Jad Naous, Michael Walfish, Antonio Nicolosi, David Mazières,
Michael Miller, and Arun Seehra, “Verifying and enforcing network paths
with icing”, COnference on emerging Networking EXperiments and
Technologies (CoNEXT ’11), December 2011, https://doi.org/10.1145/2079296.2079326
[HIST2] A. McKenzie, “INWG and the conception of the Internet: An
eyewitness account,” Annals of the History of Computing, IEEE, vol. 33,
no. 1, pp. 66–71, 2011, https://ieeexplore.ieee.org/document/5723076
[REFL] K. Neggers, “Reflections on the History and Future of the
Internet (from a technical perspective)”, Advanced Networking Guest
Lecture, University of Twente, Sep 2019, https://courses.sidnlabs.nl/anet-2019/slides/lecture1b.pdf
[Lehr] W. Lehr, D. Clark, S. Bauer, A. Berger, P. Richter, “Whither the
public Internet?”, Journal of Information Policy 9, Aug 2019, https://www.prichter.com/whither-tprc46.pdf
[DEEP] N. Foster, N. McKeown, J. Rexford, G. Parulkar, L. Peterson, and
O. Sunay, “Using deep programmability to put network owners in control”,
ACM SIGCOMM Computer Communications Review, vol. 50, no. 4, pp. 82–88,
Oct. 2020, https://dl.acm.org/doi/10.1145/3431832.3431842
[XDP] T. Hoiland-Jorgensen, J. Dangaard Brouer, D. Borkmann, J.
Fastabend, T. Herbert, D. Ahern, and D. Miller, “The eXpress Data Path:
Fast Programmable Packet Processing in the Operating System Kernel”,
CoNEXT’18, December 4–7, 2018, Heraklion, Greece, https://dl.acm.org/doi/pdf/10.1145/3281411.3281443
[MFIRST] Dipankar Raychaudhuri, Kiran Nagaraja, Arun Venkataramani,
“MobilityFirst: A Robust and Trustworthy Mobility-Centric Architecture
for the Future Internet”, ACM SIGMOBILE Mobile Computing and
Communications Review, July 2012, http://mobilityfirst.winlab.rutgers.edu/documents/mfbase.pdf
Background
Learning Outcomes
After successful completion of the course Advanced Networking (ANET) you
will be able to:
Analyze, compare, and discuss various advanced inter-networking
concepts, such as secure inter-domain routing and multi-path data
delivery.
Understand and discuss important challenges and proposed experimental
solutions, including non-IP-based internetworking systems.
Apply a domain-specific language such as P4 to implement basic data
plane functionality of an open programmable router, which is important
for future open programmable Internet infrastructures [RI].
ANET also enhances your research skills, because you will need to
independently review, analyze, summarize, and present research
papers.
Contents
Advanced concepts of IP-based internetworking
Experimental non-IP inter-networking systems such as SCION and NDN
How both handle functions such as security, mobility, multi-path
delivery
Open programmable networks as an enabler to deploy new network functions
Research challenges and new communication paradigms
Internet operations and experimental deployments (guest lectures)
Motivation
The problem that the Internet originally aimed to solve in the 1970s
was how to enable university researchers to share expensive computer
hardware via a network, which later evolved into how to make computer
networking ubiquitously available for everyone [CYCLE]. The relatively
simple design of the Internet’s core protocols (TCP/IP, BGP, and the
DNS) solved both problems beyond imagination, with the Internet now
being critical for almost every aspect of our everyday life and for our
society.
However, the success of the Internet also introduces new problems,
because it doesn’t always meet modern requirements such as society’s
need for increased security, resilience, and digital autonomy [RI] and
applications needing new network functions, such as control over and
verification of network paths [SCION] and low-latency data transfer
[CYCLE] (e.g., real-time trading or controlling swarms of autonomous
objects such as robots and cars). Researchers and engineers have
therefore developed and deployed various extensions of the Internet’s
core architecture, such as a network inspection plane [RI], routing
security [BGP1] [BGP2] and multi-path communications [MTCP1] [MTCP2].
Others are proposing and experimenting with completely new types of
inter-domain networks, for instance to support programmable and reusable
protocol functions [TROTS] [RINA1] [RINA2] or different communications
paradigms (e.g., content-centric instead of host-centric communications
[NDN]).
We offer ANET because we expect that some of these Internet
extensions and new types of internet infrastructures will have an impact
on deployed network infrastructure in the next few
years, for instance because of the widely felt decline in digital
autonomy in many societies and regions around the globe [RI]. An example
is the BGP security extensions [BGP1] [BGP2], which network operators
increasingly deploy (cf. certification-stats.ripe.net) to protect the
Internet against routing hijacks. Another example are open programmable
switches (e.g., Barefoot switches), which are already commercially
available and enable engineers to define their own packet processing
functions (e.g., for SCION or NDN traffic) in addition to IP on the same
hardware.
Objective
The objective of ANET is to enable you to understand and
evaluate advanced internetworking concepts, such as secure
inter-domain routing, multi-path communication, and path control, both
in the existing Internet as well as in alternative Internet designs
(e.g., NDN, RINA, and SCION). This will give you a wider
understanding of internetworking beyond IP-based
communications.
From the university’s perspective, ANET helps preparing the next
generation of researchers and engineers for future Internet
developments, which is important to enable the Netherlands and Europe to
stay in control of their Internet infrastructure.
Organization
ANET is a collaboration between the University of Twente and SIDN
Labs (www.sidnlabs.nl), the research arm of the operator of the .nl
top-level domain, SIDN. SIDN Labs’ goal is to increase the operational
security and resilience of our society’s Internet infrastructure, for
.nl and the Netherlands in particular.
Privacy
We’ll put the paper presentation schedule on the ANET site using only
your first name and not your last name. Please contact us if you have
any objections.
Acknowledgements
We thank our students for their feedback, which helps us further
improving ANET every year. We also thank Ralph Koning, Caspar
Schutijser, and Christian Scholten for reviewing the draft version of
this course description. Jelte Jansen kindly brushed up the layout of
the site.
