Participants of the courses "Aktuelle Themen Internet" are organizing so called "days" with a special focus on new and upcoming topics in computer science and society with a potentially disruptive force. The idea behind those events is to get students to think about possible trends and important future technologies by planning a workshop or an event. Besides the theoretical work the organization of such an event (approaching and inviting key representatives of a new technology) trains soft skills. The necessary media coverage (providing a live stream, chat, moderation etc.) leads to a media competence which is a requirement in todays workplace.
The themes for those workshops and events will be decided by the participants, work is done in groups. The following list gives some ideas for events or discussion topics. The questions for the final exam cover those topics.
The other half of the course is a weekly seminar where we discuss new technological trends and their impact on society. Participants are required to read papers and articles for each session.
If you don't like to read, don't take this course!!! There is a google drive folder where we will collect papers and other resources.
We will start with two sessions on the "digital revolution" that is currently happening and on the concept of disruption. After that we decide on further topics for discussion.
The Digital Revolution - work, technology and its consequences. See About unicorns, digital platforms and the future of technology (and society)
Disruptive Technology and the Christensen book (did not want to do this repeatedly but had to learn otherwise) Innovators Dilemma. This is mandadorty reading, because the concept of disruptive technology is so fundamental to this course. For an anti-dote to Christensen, see the critique by Jill Lepore . An alternative source:
Optional topcis for the course: (feel free to offer your own ideas). And check out the other courses of previous summer terms! Not all topics handled there are outdated!
Distributed vs. centralized: what is better? The Internet has moved between both poles frequently. Lately, centralized servers seem to have an advantage. Is that so and why? A plea for distributed can be found by Chris Dixon Why Decentralization matters. He describes the various phases on the Internet. And if we talk about distributed and Internet the Interplanetry File System as a peer-to-peer based solution cannot be left out. Murat Demirbas did a nice Review of the IPFS paper in his distributed computing course and he points out the strength and weaknesses of p2p and servers. Very good read. Some more on IPFS: From the blockchain train journal. And an excellent explanation Understanding the IPFS White Paper part 2 Dan Rayburn, Net Neutrality is a Sham. A bit unrelated but very interesting: A discussion about google reader and its follower, google news. A big difference for media creators according to Anil Dash . By the same author: The lost infrastructure of social media . Going from things we lost to things we never had: is it possible to build a distributed or federated social network? Blueprint of a distributed social network on IPFS - and its problems by Matthias Beyer is an interesting read, and so are the Comments on YCombinator. Lots of things to learn about going distributed (e.g. moderation and other experiences). Or the world's largest marketplace OpenBazaar , a fully distributed shop with bitcoins, arbiter and everything. Or take totally distributed editing with ODTRs . Collaboration without network dependencies.
Net Neutrality and attacks on the internet. Well, Trump certainly is no friend of net neutrality. Will old-style telcoms kill the googles and amazons? http://www.heise.de/ct/artikel/Klartext-1965046.html Irrtümer in Sachen Netzneutralität http://www.nzz.ch/aktuell/digital/netzneutralitaet-etno-itu-1.17231939 http://www.heise.de/thema/Netzneutralit%C3%A4t Dan Rayburn, Net Neutrality is a Sham Netzpolitik «Eine Verfassung für das Internet»http://www.nzz.ch/aktuell/international/auslandnachrichten/brasilien-kreiert-eine-internet-verfassung-1.18272165
I have pulled this to the top because of current developments in the US. Looks like big old AT&T will be great again...
Load balancing and komplex protocols like gRPC. Always a topic for internet services. Using Envoy to Load Balance gRPC Traffic by Mike White gives you some hints on why balancing http/2 ain't easy. In case you are looking for more basic information on LB Introduction to modern network load balancing and proxying by Matt Klein should be on your reading list.
Current developments and trends in the Internet itself. Is the Internet still a disruptive force? What are the effects of IPV6 and will it solve the symmetry problems? This topic is tightly related to "distributed web" and "net neutrality".Geoff Huston on Adressing 2017 is an excellent source for internet topics like NATing, DNS, adressing, BGP and other core technologies the internet is riding on.
The Internet of the future: A big private Content Delivery Network? Geoff Houston, The Death of Transit and Beyond and Alan Mauldin adds additional information on submarine cables etc. Looks like the Internet is getting restructured and privatized even more. A good explanation and overview in German from Monika Ermert, Heise.de
Faster than TCP? Let's take a look at The QUIC Transport Protocol:Design and Internet-Scale Deployment by Adam Langley, Alistair Riddoch and others from Google. What makes the protocol faster than TCP?. Is http/2 faster or does it have better throughput? And what happens in case of Network troubles?
Blockchain has been all the rage lately. A sore spot in most protocols is the energy intensive way proof-of-work is done. Is there a reliable alternative like proof-of-stake? Next Generation Decentralised Ethereum Proof of Stake Pool .Or even some consensus protocol? And Do you need a blockchain?. Stellar is an internet level, open membership, byzantine consensus protocol. so there is more than blockchain...
Well, there is no denying that AI and machine learning are having a major impact on the world. How will they affect software development and internet services? Ambient AI Is About to Devour the Software Industry shows how Amazon ties cloud and AI technologies together into one seamless dashboard for developers. Other alternatives like The Coming Software Apocalypse -A small group of programmers wants to change how we code—before catastrophe strikes. place a bet on formal methods to save software development.Machine Lerning Driven Programming
You want your new service to get really BIG. How do you get it e.g. on Google's application engine? How to Lift-and-Shift a Line of Business Application onto Google Cloud Platform by Andy Wu, Solutions Architect, Magenic, gives a nice example. And while we are looking at that, we might also take a look at how Netflix uses the cloud to become the number one video on demand corporation.Netflix: What happens when you press play?
If the future belongs to the cloud - what is YOUR future? . System engineering and software development between AWS and neuronal nets.
Strangeloop 2017: the latest on Technology. This is one way to make sure you don't miss important developments.
Internet access in times of trouble. How to use IPFS to circumvent censorship.Catalonia Referendum Voting
Serverless in the cloud and at the edge and everybody is a programmer. Let's take a look at future computing topologies and how services are tied together by regular people.. A new paper on serverless by Rob Gruhl on Scaling Serverless
Biocomputing is getting hotter and hotter. Last year we had a talk on 3D-printing of organs. Should we get one on CRISPER/CASS?
High frequency trading systems: When a Microsecond Is an EternityCarl Cook explains high performance techniques and why they are so important in our world today.
How To Make It In The Music Business . Another business that would not be possible without the Internet. Any new developments here? There is a lot of collaboration going on .Intesting, how streaming changes album sizes!
Internet Security: Template injection attacks. A nice example for a subtle attack on servers.
Big data meets Big Brother as China moves to rate its citizens, The Chinese government plans to launch its Social Credit System in 2020. The aim? To judge the trustworthiness – or otherwise – of its 1.3 billion residents, By Rachel Botsman, 21 Oct 2017 wired magazine.Die Big-Data-Diktatur
Laying REST to rest? Is it time now to retire REST APIs? The new GraphQL approach might just do that for certain classes of applications. Eric Baer, a GraphQL Primer Part 1 and 2).You can learn more about GraphQL in the Tutorial
Scalability is still king of the Internet. Kobi Hikri, Get Rid of that Bottleneck describes queuing techniques for internet facing services and applications. A nice example of a high-throughput process pipeline
Blockchains from a Distributed Computing Perspective, MAURICE HERLIHY, Brown University. What can we learn about BC by comparing it with regular DS techniques?
LaunchDarkly: How to test in Production. Videos. Does not sound very sexy, but how do you deal with DLNNs, microservices etc. in production?
If you believe Harari, the future of humans is in bioengineering (whatever human will mean after this). Bert Hubert's DNA seen through the eyes of a coder can help you understand the basics which are - surprise - digital. An interesting comparison of codes, alphabets and sentences in computer science and biology. A 2 hour presentation is also available.. This could be a perfect primer for a Day on CRISPER/CASS. And remember: the guys who built the 3D-organ printer last summer were web engineers, not biologists.... More on bio and computing on ycombinator and Brainless Embryos Suggest Bioelectricity Guides Growth . Are bio-design patterns really this much different from regular software design and architecture?
API design and usability. An extremely important and almost always forgotten topic. That is why we have to live with creepy stuff like a socket API. Five API Usability Lessons from Flutter (DartConf 2018) by Tao Dong has some nice tips. Look at the conceptual model diagram to capture the users model! Let me give you some extra advice: Let the users of your API design the interfaces and create test programs using your API first. This will prevent stuff like the socket API. If you are into programming languages you should also look at linear and behavioral types which can make the sequential use of APIs safe.
A nice example of integrating Deep Learning Services into an application. A scalable Keras + deep learning REST API by Adrian Rosebrock.
I am sure all of you are testing their web apps seriously. But do you use fuzzing to detect the unexpected? The art of fuzzing by Rene Feingruber is a nice introduction to this technology.
The Internet seems to be a breeding place for all kinds of bubbles and hype. Bitcoins, fake news, speculation and so on. But is this really something new? The Dutch Tulip Mania: The Social Foundations of a Financial Bubble describes a bubble in the 17th century in Holland. A. Maurits van der Veen investigated how a single tulip came to be as costly as a house. Any patterns to be found here?
How to make a dynamic site static with the help of a CDN. Leonardo Losoviz is the creator of PoP, an open source framework for building decentralized social networks. HQ Trivia suffered severe outages through Super Bowl - something that has to do with a lack of edge data and processing.
On the importance and difficulty of memory safety in languages: Memory Safety in Rust:A Case Study with C. Do you find the problems in the C code?.
Webhooks or the ability to collect real-time data from IoT devices is becoming a critical feature. The article on medim argues against the use of servers and favors a FAAS archictecture. Richard Moot, Stop Using Servers to Handle Webhooks. How does such an architecture look like?
The future of digital media seems to lead straight into fake-news, computational manipulation and human puppets. Generative adversarial networks can manipulate audio/video/image content to create believable "post-facts". This raises questions about the common information layer that keeps societies together and which has been traditionally maintained by mass media. Charlie Warzel, buzzfeed.com news reporter on Aviv Ovadya . In the same vein: Web Literacy for Student Fact Checkers by Michael A. Caulfield
With retargeting being more and more perfect you should probably start using TOR more frequently to avoid personalized pricing etc. How does TOR work?
Messaging is all the rage nowadays with whatsapp, slack and others. But how does e.g. Slack perform in development teams?. A critical article. "slack empowers your worst people to overwhelm your best. It has that in common with the open office. Goodhart’s law is the one where turning a metric into a target makes it a bad metric. Slack subverts valuable work by making productivity = availability on slack."
The Future Today Institute has published 225 trends for the future. AI, scurity and much more. The list of contents alone lets us speculate about cross-trend developments which could lead to even greater disruptions. download the report.
Interaction Design is getting more important than ever. When Paul Sonnentag returned from his off-site term, he told me about Bret Victor's ideas and we started a project for the summer term. His homepage is very interesting and there are lots of videos showing his projects.
The internet has given us massive scale data processing. Want to know how this evolved at Google? Tyler Akidau's presentation
Critical Infrastructure: Patterns and Anti-Patterns of Secure Systems. The internet exposes critical services to the whole world. What can/must be done to keep those services secure? To avoid getting blackmailed by trojans? I gave this talk at an electrical engineering conference and it fit perfectly to our course. Robust Systems
Arbeiten 4.0: Corporate Culture in Internet Times. The fast progress of internet companies requires a different corporate culture which values speed and independence over other things. We are looking at two typical representatives: Reed Hastings, Netflix Culture - Freedom & Responsibility, and (Henrik Kniberg & Anders Ivarsson, Scaling Agile at Spotify ) and see how social and technical organization of work interleaves. We will learn about Conways law and why Netflix believes it should be like Bayern München. One more hint is having a bench strategy:. On how to create an engineering team and how google does it. Further tips: the topic has been talked about at QCON and also Heise/Gunter Dueck gave a comment. And last but not least a new feature from Dueck on "Arbeiten 4.0" . This topic fits nicely to our Entrepreneur-Day later (see the schedule below). But ask yourself: who is interested in seing start-ups all over the place and why?
Microservices and Lambda Architecture. An example for a microservices approach is given by netflix (Adrian Cockcroft GOTO Berlin - November 2014, Migrating to Microservices ) . A good intro comes from Lewis/Fowler, (Microservices ). Practical insight comes from Kevin Scaldeferri's talk at OSCON 2015 (CONTINUOUS DELIVERY AND LARGE MICROSERVICE ARCHITECTURES, Reflections on IonCannon. And finally a free book on Microservices from the NGINX guys. A good intro to serverless computing can be found at Amazon. You should at least understand the architecture examples given in the getting started doc. For those who want more: Martin Fowler talks frequently about it: . And for the friends of IBM
We did a thing on Bitcoins two summers ago. Last summer, the block-chain seemed to be the hottest thing in finance and there are lots of ideas on how to use it for other things as well. Unfortunately, there are quite severe scaling problems behind the protocol. An excellent book by Ed Felten et.al. allows us a good look at the internals of a working blockchain and its spin-offs (like name-coins etc.). Get it from here. I found the rat race for ever faster mining hardware an interesting case. We could discuss alternatives for the proof process which are not based on burning excessive amounts of energy ("proof of stake"). And the programming language integrated in Ethereum and its impact on smart contracts. Interesting mesh between distributed web and blockchain tech: A Decentralized Content Registry for the Decentralized Web
How do we find information? How the latest important papers? Where do we go for technologies? Which sites do we HAVE to read? Frequently I get asked about my sources of information and I think we should just throw together what we do in a best practice session. You should know at least one portal, one paper site and one conference for your special areas each. Some examples for good sources in distributed computing are HighScalability.com or infoq.com as a portal (read weekly updates). The famous "Morning paper" site by Adrian Colyer and the QCON Conference in London and SF, see a good writeup by Andy Butcher . For the hardware fan: mechanicalsympathy.com
The De-Centralized Web. Decentralized solutions - no longer viable? A vision or an illusion? We will talk about Brewster Kahle and his vision of a decentralized and free web and internet. What are the things we would need for this? What kind of technology is here to help us? Brewster Kahle's talk . Some technology parts: Named Data Networking Ipfs: - a good video on this page, Namecoin , a facebook alternative? Safebook (the interesting Matroshka design pattern), finally: what internet providers (the telkoms) know about us. And hot from the conference. And for some practical ideas on disrupting the big digital platforms: What happened to Austin, TX, after Uber and Lyft left town: . I guess the main thing here is the problem behind Man-in-the-middle like business models and whether they can be replaced with peer-to-peer approaches. See: Disrupting Uber, Driver-owned apps could end Uber’s exploitative reign over the ride-share market. by Vic Vaiana
Container Technology and Unikernels. One of the hottest trends in computing right now. Here is the talk from Kleindienst/Frey And go and look up some info on Rumpkernel or MirageOS. A must read: will containers replace hypervisors What is happening here? What could be the end of the line for some time? See: The answer is dynamic code generation instead of re-packaging huge modules!
Algorithms are no less disruptive than hardware-based technical revolutions. You should know at least some from the following groups (short descriptions can be found on wikipedia): Probabilistic algorithms/data structures (sketches). They work e.g. by observing bits in hashes. Bloomfilter, hyperloglog, count-min belong there. Another group deals with lock free algorithsms. LMAX is a wonderful example here, but there are many others too. A third group is ultra-fast algorithms running in CPU caches. You should know one and understand the differences between batch/online/one-shot Algorithms. Another group deals with scalability and provides extreme parallel processing capabilities. Map/Reduce is one example. We dealt with security in the blockchain technology already and you should be able to explain how it works and how it can be used. Another topic: algorithm based feeds And finally: distributed consensus, as it was mentioned in the course: DS consensus with Paxos explained. Fascinating: special algorithms for high-frequency trading
Finally, another potentially disruptive development: Smart Home and Internet of Things. The slides are on the google drive. For a security background on the problems Smart Home is facing: ENISA report on smart home security . How to program the IoT: How Lil Todo Syncs Tasks Across Multiple Devices Just Using Dropbox
The million things we did not talk about yet in this course:
Current developments and trends in the Internet itself. Is the Internet still a disruptive force? What are the effects of IPV6 and will it solve the symmetry problems? This topic is tightly related to "distributed web" and "net neutrality".Geoff Huston on Adressing 2016 is an excellent source for internet topics like DNS, adressing, BGP and other core technologies the internet is riding on. Is peer-to-peer the future of the Internet? see: Peer gewinnt, 28.08.2015 – Niels Boeing, Claudia Wessling, Technology Review.
Welcome to the matrix - the internet of things is about to become real - or?. http://www.heise.de/tr/artikel/Willkommen-in-der-Matrix-2551024.html
Scalability, performance and costs of running large scale internet sites: does it make sense to move away from Amazons cloud? What is feedback control and how does Netflix handle it? The slides from Uwe Friedrichsen, codecentric.de. highscalability.com How Github concquered everybody (network effects etc.) http://www.wired.com/2015/03/github-conquered-google-microsoft-everyone-else/ and of course: microservices: http://www.indix.com/blog/microservices-what-are-they and an exciting distributed search architecture with dist. consensus and anycast DNS: http://highscalability.com/blog/2015/3/9/the-architecture-of-algolias-distributed-search-network.html. There are three different types of internet traffic: regular, unplanned spikes and planned spikes (like Olympic games). How Facebook live video deals with those cases.
Marketing in internet times: 30 Billion ads per day! http://highscalability.com/blog/2015/3/9/applovin-marketing-to-mobile-consumers-worldwide-by-processi.html
Love in the Internet Age: the user experience with online dating algorithms and sites.http://www.zeit.de/digital/internet/2015-02/tinder-online-dating-hacks-erfolg
Cyborgs - from smartphone to wearable to implants: Internet driven enhancements to humanity. Let's develop a vision on google glasses, embedded sensors and actors and a new definition of where your body ends...http://www.nzz.ch/feuilleton/unter-die-haut-1.18485958 http://www.futuremag.de/interview/mensch-und-maschine-schon-heute-verschmolzen-interview-mit-stefan
Corporate Cultures of Unicorn Companies - How is it tied to continuous delivers, lean enterprise and agile development?
Messaging: the new paradigm. Will Slack etc. change the way we shop?
Crowds and Crowd-Funding -and when does it fail? http://www.heise.de/newsticker/meldung/Meilenstein-fuer-Crowdfunding-Plattform-Kickstarter-Insgesamt-eine-Milliarde-Dollar-zugesagt-2131171.html
Getting rid of information: Snapchat etc. http://www.heise.de/tp/news/Weg-mit-den-Bytes-2121769.html
Microservices - the architecture for scalable Internet Services
Darknets and Anonymity: http://www.heise.de/tp/artikel/39/39575/1.html and the getting started pages of the torproject.org
Router Attacks/ Anatomy of a DDOS http://www.heise.de/security/meldung/Grossangriff-auf-Router-DNS-Einstellungen-manipuliert-2132674.html
Internet Resilience: ENISA Study
Thinking Machines: https://www.edge.org/conversation/john_brockman-what-to-think-about-machines-that-think There seems to be a growing concern about humans losing to machines in the end. Anybody playing GO?
Neuroscience and Computer Science: I think, therefore I heal: the weird science of neurofeedback theverge.com http://m.heise.de/tr/artikel/Schrittmacher-fuer-das-Gedaechtnis-2092798.html?from-classic=1 http://www.kurzweilai.net/wearable-neurocam-records-scenes-when-it-detects-user-interest http://www.kurzweilai.net/neuroscape-lab-visualizes-live-brain-functions-using-dramatic-images
Inhouse-Navigation, Apple I-Bacon, Android
Eli Pariser, the filter bubble (and perhaps the price bubble..). On "echo chambers" and how the topology of networks can create a false feeling of majority.
The Master Switch - Rise and Fall of Information Empires, Tim Wu
Campact und Co.: Demokratie durchs Internet - ein Erfolgsmodell? https://www.campact.de/netzneutralitaet/appell/empfehlen/
Internet Payment: http://www.heise.de/tr/artikel/Kreativer-Zerstoerer-2127847.html
Mesh Networks, open WLAN www.heise.de/tr/artikel/Die-Ruechttp://www.heise.de/tr/artikel/Die-Rueckeroberung-des-Internets-2127869.html?view=printkeroberung-des-Internets-2127869.html?view=print and http://www.heise.de/download/firechat-1194036.html
Dave Eggers, the circle: The Circle is the exhilarating new novel from Dave Eggers, best-selling author of A Hologram for the King, a finalist for the National Book Award.
The Algorithmic Age: The positive and negative sides of modern algorithms. We could start with Code-Dependent: the pros and cons of the Algorithmic age, by Lee Rainie and Janna Anderson and then tackle the famous Cathy O’Neil, author of Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy, who pointed out that predictive analytics based on algorithms tend to punish the poor, using algorithmic hiring practices as an example.
More in that vein: “Why Should I Trust You? Explaining the Predictions of Any Classifier Ribeiro et al., KDD 2016, Explaining Outputs in Modern Data Analytics, Zaheer Chothia, John Liagouris, Frank McSherry, Timothy Roscoe Systems Group, Department of Computer Science, ETH Zürich, How the machine "thinks": Understanding opacity in machine learning algorithms, Big Data & Society January–June 2016: 1–12, Jenna Burrell, European Union regulations on algorithmic decision-making, and a “right to explanation”, Bryce Goodman, 1∗ Seth Flaxman. Now, this could easily form the basics for a very interesting "Digital Rights Day" with my colleague Prof. Veddern. see also: Algorithmic collusion and price-fixing January 9, 2017 Cathy O'Neil,
The brain and its computer. Should we model computations like the human brain? Well, do we even know how the brain works? How the Brain Might Work: Statistics Flowing in Redundant Population Codes, Xaq Pitkow 1,2 and Dora E Angelaki 1,2
Content Delivery Networks. Best Practices For Using A Multi-CDN Strategy: How To Balance, Prioritize and Optimize Traffic, Dan Rayburn
A critique of Machines: All Watched Over by Machines of Loving Grace 1h | Documentary | TV Mini-Series (2011– ) Episode Guide 3 episodes All Watched Over by Machines of Loving Grace Poster A series of films about how humans have been colonized by the machines we have built. Although we don't realize it, the way we see everything in the world today is through the eyes of the computers.
Engineering privacy on the internet: Engineering Privacy for Your Users, Session 709 Jessie Pease Privacy Engineering Julien Freudiger Privacy Engineering and: ANONYMIZATION AND RISK, Ira S. Rubinstein * & Woodrow Hartzog **
facebook and disaggregated networking: new approaches in networking: https://code.facebook.com/posts/1887543398133443/disaggregate-networking-recap/
Scalability in the internet age: The Infrastructure Behind Twitter: Scale Thursday, January 19, 2017 | By Mazdak Hashemi (@mazdakh), VP of Infrastructure and Operations Engineering [19:35 UTC]
Here are the dates for our days and special guests.
13.4.2018, Big Data Research at ETH Zurich, Prof. Karten Borgwardt at Uni Esslingen. 16.00 with apero.
20.4. "Industry Meets Interaction Design", Joachim Charzinski
Blockchain Technologies? Charzinski with Alumni:
Please watch this page closeley for changes in dates or topics. Most speakers work in the industry and sometimes need a change in schedule due to that fact.