18-847: Data-intensive and Cloud Computing and Storage (Fall 2010)

Data-intensive computing and cloud computing have become important forms of computing, and both appear poised to grow into dominant roles. Data-intensive computing (DISC) refers to analysis and information extraction from large and sometimes dynamic data corpi. Cloud computing refers to shared (multi-tenant) use of third-party computing and storage resources (and sometimes software setups) in place of dedicated resources. Each is interesting on its own, and the use of cloud computing for data-intensive computing is both inevitable and critical.

In this course, we will explore the state-of-the-art and research directions relating to data-intensive computing and cloud computing. Included in this scope will be case studies of existing systems, compute and storage architectures, programming models, middleware and building blocks, and administration/automation. In our discussions, we will explore various metrics of goodness for alternate approaches, including efficiency, performance, robustness, complexity, ease-of-use, and so on.


  • In-class presentation and discussion of seminal and timely research papers and articles. Each student will be expected to read all papers, participate in the discussions of them, and present and lead the discussion on several of them.
  • Guest lectures by experts in the field
  • Semester-long research project
  • Readings

    Each class meeting will have required readings, to be discussed in class. Usually, these readings will consist of relevant technical papers or articles. The readings listed should be read BEFORE class on the assigned day. Otherwise, how could you participate in the discussion?

    For each required reading, there will be a pre-designated discussion lead, who should prepare for that role ahead of time. Everyone is assumed to be reading the papers before class, so the discussion lead is not "presenting" the paper to an uninformed audience. Rather, the discussion lead should focus on sharing some of their insights to shape the discussion. It is fine to use projected slides to help in playing the role of discussion lead, but it is not required. Basic topics worth trying to cover in leading the discussion include:

  • Very brief reminder of the overall focus of the paper/system, just to focus everyone
  • Quick reminder of the architecture/design, if the paper presents or refines a system
  • Primary lessons learned (by you) from the paper. Note that these may or may not match those identified by the authors.
  • Most interesting insights you got from the paper, which again may or may not match those highlighted by the authors
  • Context of how the paper fits in with related work, in your view
  • Questions left open by the paper and doubts raised in your mind upon reading it
  • Any flaws or shortcomings that you noted
  • Don't worry that you might have different answers, on some of these, than others (including the instructor)... some of these are opinion-based and will be the source of debate during our discussions.

    See the Readings page for the list of papers and associated dates for discussion.


    This class is intended for those pursuing research in its topic space. As such, it requires Ph.D. student status or permission of the instructor. Permission might be given to non-Ph.D. students, for example, who have taken and done well in 15-712 or who have demonstrated research skills in other ways. Students without substantial experience reading research papers will struggle, as will those without sufficient time to read and thoroughly comtemplate several papers every week.

    Mobile devices and other distractions

    Research on learning shows that unexpected noises and movement automatically divert and capture people's attention, which means you are affecting everyone's learning experience if your cell phone, pager, laptop, etc. makes noise or is visually distracting during class. For this reason, we allow you to take notes on your laptop, but insist that you turn the sound off so that you do not disrupt other students' learning. If you are doing anything other than taking notes on your laptop, please sit in the back row so that other students are not distracted by your screen.

    No recording of class meetings

    No student may record or tape any classroom activity without the express written consent of all instructor(s). If a student believes that he/she is disabled and needs to record or tape classroom activities, he/she should contact the Office of Equal Opportunity Services, Disability Resources to request an appropriate accommodation.

    Final Note

    * Everything here is subject to change.