SIPB Cluedump Series 2016
From SIPB Cluedumps
SIPB Cluedumps are informal technical talks open to the entire MIT community. They cover topics that are of general interest, such as web browsers, and topics specifically for the MIT computing community, such as Zephyr and Scripts. Cluedumps are usually one to two hours long, and we provide snacks.
If you would like to receive weekly announcements about Cluedumps, add yourself to firstname.lastname@example.org or email email@example.com.
For more information or if you'd like to give a Cluedump, please contact the organizers at firstname.lastname@example.org.
|Date: March 26, 2016, at 3:00 PM|
|Presenters: Will Vahle and Thomas Delgado|
| Abstract: Using software-defined radio (SDR), a single piece of hardware can be rapidly repurposed into a wide variety of very different sorts of radio receivers without having to build and configure new hardware for each new band, modulation strategy, or error-correction scheme.
Join us as Will Vahle and Thomas Delgado show off some inexpensive, commercially-available hardware (the RTL-SDR) and talk about the toolchain required to use it. They'll demonstrate using SDR to listen to aircraft, maritime, and police radio, as well as FM broadcast stations. Also included: Van Eck and TEMPEST side-channel attacks, as well as receiving (but not decoding) cellular signals.
|Date: March 30, 2016, at 7:00 PM|
|Presenters: Alex Chernyakhovsky|
| Abstract: Alex works in Google's Content Delivery Network, and can discuss architecting clusters; load-balancing traffic; job scheduling; building distributed, replicated, highly available systems; operational considerations; how to achieve reliability; high-performance computing; and many related topics.
This will be an interactive, question-driven talk. Audience interaction is highly encouraged, and snacks will be served!
|Date: April 14, 2016, at 7:00 PM|
|Presenters: Samuel Dukhovni|
| Abstract: Sick and tired of the NSA reading all of your emails? Samuel will walk through setting up PGP in Thunderbird, so you can have your top-secret communications encrypted end-to-end. Let's all pretend that not everything they need is in the metadata anyway! :D
Snacks will be served. Your taxes can wait.
|Date: April 27, 2016, at 7:00 PM|
|Presenters: Alex Rosenblat|
| Abstract: Uber manages a large, disaggregated workforce that delivers a relatively standardized experience to passengers while simultaneously promoting drivers as entrepreneurs whose work is characterized by freedom, flexibility, and independence. Uber, like other companies in the on-demand economy, uses its identity as a platform and a technology company to restructure its employment relationship to drivers, who are classified as independent contractors. It claims to provide a "lead generation application" for drivers to connect with passengers, but this neutral branding of its role as an intermediary belies the important employment structures and hierarchies that emerge through its software application. Through a 9-month empirical research study of Uber driver experiences, myself and my colleague, Luke Stark (NYU), found that Uber leverages significant control over how drivers do their jobs, but this control is structured to be indirect. The opacity and efficacy of control is achieved through a range of semi-automated managerial functions, but foremost amongst these are: algorithmic labor logistics management; driver surveillance and the rating system; and performance targets and policies that limit the choices drivers can make to optimize their individual earnings on the system. For a quick synopsis, see media coverage from The Awl, WSJ, MIT Technology Review, or an article I wrote for HBR.
How can you help?
I'd like the hive mind to help think through the privacy considerations in protecting the identities of drivers in future research, and to examine the ways that platforms can leverage user data (drivers and passengers) to create targeted prices and tiered wages. I'm also interested in issues of user design and deception (see this article I wrote for Motherboard about Uber's phantom cars), and the potential for automating inequities or automating power and knowledge symmetries between workers and platforms. There are broader questions I'd like to dig into as well, such as: When technology creates new efficiencies and reduces friction in transactions between demand (customers) and supply (workers), where do latent points of friction emerge? How do the rhetorics of marketplace efficiency clash with the goals of individual earners? (Example: tipping is "inefficient.")
|Bio: Alex Rosenblat is a researcher and technical writer for the Intelligence & Autonomy Initiative at Data & Society, a project supported by the John D. and Catherine T. MacArthur Foundation. She tweets @mawnikr.|
|Date: May 4, 2016, at 8:00 PM|
|Presenters: Jonathan Harvey-Buschel|
|Notes: Slides in PDF or LibereOffice/OpenDocument Presentation formats|
| Abstract: Since the early 80's, the concept of anonymous electronic cash has been an unsolved problem amongst the cryptography community. While Bitcoin is far from a perfect or proven implementation of this concept, it has been running in the wild for over 7 years and has grown from a 9-page whitepaper into a multi-billion dollar network supported by thousands of full nodes and the consumption of hundreds of megawatts.
To understand some key properties of the network, we will first cover the primitives used in Bitcoin and how they relate to various aspects of participating in the network. Given this background, patterns observed in the real world can be related to parts of the protocol. With an understanding of the current state of the network, we will explore the challenges facing Bitcoin today, and some proposed solutions.
Snacks will be served.
|Date: May 25, 2016, at 7:00 PM|
|Presenters: Colby Gutierrez-Kraybill|
|Notes: Presentation slides and related paper|
|Abstract: The Allen Telescope Array (ATA) has been producing science data since 2007 at the Hat Creek Radio Observatory (HCRO). The instrument was noted as a significant path finder for technologies that are currently being incorporated into the Square Kilometer Array designs. Built to aggressively save on costs for the front end antenna design with multiple signal paths, the ATA was designed to have different generations of processing backends, including large scale computer clusters as well as purpose built FX correlators, beamforming elements, and one-off experimental computational systems. Originally designed and built by the Radio Astronomy Laboratory at the University of California, Berkeley, working with the SETI Institute, the instrument is now operated by SRI International and the SETI Institute.|
|Bio: Colby Gutierrez-Kraybill is a software engineer and systems administrator. He was part of the original team that built the ATA, working as the lead software developer for the Radio Astronomy Laboratory at UC Berkeley from 2001-2011. From 2011-2012 he was the onsite operations manager at the Hat Creek Radio Observatory for SRI International. He and his family moved to HCRO in 2008 and lived on site until 2012. He now works at the MIT Media Lab on large scale web based systems for the Lifelong Kindergarten and NeCSys groups.|
|Date: July 27, 2016, at 7:00 PM|
|Presenters: Ray Hua Wu|
|Notes: Standard world time zones, Solar time vs standard time, Time in Australia, Time zones of the world|
|Abstract: This talk will be about how unexpectedly complicated they are and why if anyone ever needs someone to code up a time zone system you should not volunteer.|
|Date: September 14, 2016, at 7:00 PM|
|Presenters: Ray Hua Wu|
|Abstract: Athena is MIT's computing environment, and learning how to use it can help anyone at MIT. Find out how to discover information about other people; manage mailing lists; and how to get into Athena's computer clusters if you've forgotten the door combo. Also, learn about ways to use the Athena environment when you're not at an Athena machine.|
|Date: September 27, 2016, at 7:00 PM|
|Presenters: Dalton Hubble and Colin Hom|
| Abstract: |
Provisioning CoreOS and Kubernetes on Hardware (Dalton Hubble)
Kubernetes is a powerful system for operating application containers across a cluster of machines. In this talk, we'll explore CoreOS cluster provisioning and Kubernetes setup on hardware. To start, we'll cover PXE network setup and Ignition, CoreOS's built-in early-boot provisioning tool. Then we'll discuss bootcfg, a service which matches machines to profiles to provision complete clusters. We'll walk through PXE booting machines, installation to disk, and automated provisioning of a multi-node etcd key-value store and multi-node Kubernetes cluster. We'll show how the approach extends across machines and to provisioning many different kinds of reference clusters we work on.
Kubernetes on AWS (Colin Hom)
Kubernetes and AWS go together like peanut butter and jelly - both are popular, and combining them just makes sense. The talk begins with an overview of how Kubernetes is deployed on and integrates with the AWS platform. We'll cover diverse topics including DNS zones, load-balancers, and route tables.
Building on Dalton's talk about deploying Kubernetes on physical hardware, we'll explore the challenges associated with extending the Kubernetes abstraction across datacenters and cloud platforms. We'll discuss cutting-edge features and design patterns emerging in the Kubernetes ecosystem. To conclude, I'll ask the audience to speculate on how the movement towards containerized applications and orchestration frameworks will change the industry.
| Bio: |
Dalton Hubble is a software engineer at CoreOS. He builds services and apps for provisioning hardware into clusters and works on Tectonic, CoreOS' enterprise offering. He was formerly an engineer at Twitter and studied at MIT.
Colin Hom is an infrastructure engineer at CoreOS. CoreOS is the company delivering Google's Infrastructure for Everyone Else (#GIFEE) and running the world's containers securely on CoreOS Linux, Tectonic and Quay.