Due Tuesday, September 22: Nodes and Networks of Computation
In previous weeks, we studied three components of a traditional stand-alone computer system: Input (Data) → Processing (Algorithms) → Output (Visualizations). This week, we ask what changes when a single computer becomes one node in a much larger network of computers. We will return to human and social networks later in the semester; for now, our focus is on computational networks and the people, institutions, and values built into them.
Connecting computers allows data, memory, and computation to be distributed across many machines. A network may continue operating when an individual node fails, but resilience is not automatic: hubs, bottlenecks, shared infrastructure, and design choices can also create new dependencies and points of failure. Networks can acquire capacities that no individual node possesses. We will use the idea of emergence to investigate how these collective capacities arise, and network effects to consider how the usefulness or power of some networks changes as participation grows.
Networks now connect not only computers and phones but sensors, vehicles, appliances, industrial systems, and other physical objects. This Internet of Things can extend perception and coordination across enormous distances while also extending surveillance, vulnerability, and control. Encryption and blockchain offer another way to explore networked trust: how can people exchange information or maintain a shared record without entrusting every decision to a single central authority? Blockchain does not eliminate trust; it asks us to consider where trust and power move when they are placed in protocols, code, infrastructure, and collective agreement.
Throughout the week, keep our central question in mind: When does a network acquire capacities—and forms of power—that none of its individual members possesses?
These foundations will prepare us for our conversations later in the week about surveillance, hacking, cybersecurity, cyberwarfare, anonymity, censorship, and network governance. The same connections that allow networks to distribute information, computation, and trust can also create new opportunities for monitoring, attack, disruption, and control. As you complete Tuesday's homework, begin noticing not only what each network makes possible, but also who can see it, influence it, interrupt it, or govern it.
Homework assignments
DataCamp and the visualization mini-project
- Complete Understanding Data Science, Chapter 3: Preparation, Exploration, and Visualization.
- There was a mismatch between the assignments listed in DataCamp and those listed on the course webpage. Please check your DataCamp record and make sure that Chapters 1–3 of Understanding Data Science are complete.
- If you have already completed Chapter 4, don't worry: that work was not wasted. We will return to its material on experimentation and prediction later in the semester. For Tuesday, simply make sure that Chapters 1–3 are complete.
- If you have already completed all three chapters, use this time to work on your visualization mini-project.
- As you complete Chapter 3, connect its treatment of structured and tidy data, missing values, outliers, exploratory analysis, and dashboards to the decisions you are making in your own project.
Visualization Mini-Project #1 is due Friday, September 25, in the evening. We will spend time on the projects in Tuesday's class.
How computers become networks
- Inside a Google Data Center
- How Computers Work: Hardware and Software (5:22), Code.org, January 2018
- The Internet: IP Addresses and DNS (6:44), Code.org, September 2015
- The Internet: Packets, Routing and Reliability (6:25), Code.org, September 2015
Networks, emergence, and the Internet of Things
- Emergence: How Stupid Things Become Smart Together (7:30), Kurzgesagt, November 2017
- Network Theory Overview (5:50). This and the following video provide a brief introduction; we will study graph theory more deeply during our later week on social networks.
- Graph Theory Overview (4:21)
- The Wired Guide to the Internet of Things, Arielle Pardes, Wired, November 2020. Read for its explanation of connected devices and sensors; some examples and forecasts reflect the time when it was published.
- The Internet of Things: A Primer, Information Is Beautiful. Click through the diagrams in this interactive visualization.
Encryption, blockchain, and trust
- Asymmetric Encryption (4:40), Simply Explained, October 2017
- What Is Blockchain? (6:26), Centre for International Governance Innovation, January 2018. Use this as an introduction to the mechanism and its ambitions, not as a prediction about blockchain's future importance.
Questions to prepare for class
Come prepared to answer these questions from the homework. Short notes are welcome; polished written responses are not required.
- How do IP addresses, DNS, packets, and routing work together to move information across the internet?
- What can a network accomplish that an individual computer cannot? Give one example of an emergent capacity.
- What are the benefits and risks of connecting ordinary objects through the Internet of Things?
- How does asymmetric encryption allow people to communicate or establish trust without sharing a private key?
- What problem is blockchain designed to solve, and how does its approach differ from keeping records through one central authority?
- From DataCamp Chapter 3: what is one way that missing, messy, or poorly structured data could affect your visualization mini-project?
Terms you have probably encountered before
Be prepared to explain these in your own words:
- hardware and software
- operating system
- internet and network
- server
- IP address and domain name
- Internet of Things
- encryption
- missing data and outlier
Terms that may be new or less familiar
Use the assigned materials to develop a working understanding of these terms and bring questions about anything that remains unclear. You do not need to memorize formal definitions.
- node and edge
- protocol
- DNS
- packet and routing
- bandwidth
- redundancy and distributed system
- centralized and decentralized network
- designed and random network
- emergence and network effect
- symmetric and asymmetric encryption
- public key and private key
- cryptographic hash
- blockchain and consensus
- structured data, tidy data, and exploratory data analysis
Questions we will explore together in class
- When does a collection of connected computers become something more than the sum of its individual machines?
- Is the internet decentralized? In what ways is it decentralized, and where is it highly centralized?
- Does connecting more objects make a system smarter, more vulnerable, or both?
- Does blockchain eliminate the need for trust, or relocate trust into code, protocols, developers, infrastructure, and collective agreement?
- Is computational consensus the same thing as social or political legitimacy?
- Who gets to design a network's rules, and whose interests do those rules serve?
- How have the collection, structure, and limitations of your dataset constrained the story you can tell?
Coming next: Thursday's homework will introduce the Unix shell in preparation for exploring a live network with commands such as ping, nslookup, and traceroute/tracert in class.