Spring 2016: Datacenter Architecture
ECE 590, TuTh 3:00-4:20PM, Hudson 208

[Syllabus]
This course covers advanced topics in data centers with an emphasis on computer architecture and systems. This course surveys recent advances in processor, memory, network, and storage. And it surveys modern software systems that run in computing clouds. Discussion-oriented classes focus on in-depth analysis of readings. Students will learn to reason about datacenter performance and energy efficiency. Students will complete a collaborate research project.



Project

The project is open-ended and students may form groups of up to three to pursue a project related to computer architecture and distributed systems. Deliverables are scheduled for the remainder of the term.

[Project Proposal]



Readings

Participation/Discussion. This course uses a seminar, not a lecture, format. Each class covers particular topics from assigned readings. Students are expected to read the assigned papers and to prepare for course discussions. A student will be assigned to lead the discussion for each paper.

Response Papers. The students should prepare an insightful critique of the assigned papers due at the beginning of class. These response papers should take the form of a constructive paper review, including (1) summary, (2) strengths, (3) weaknesses, (4) directions for future work. These response papers should be no longer than one page per class. Papers will be evaluated for brevity and depth of insight.

[Sample]

Discussion Lead. Paper leader should prepare pose questions to stimulate a discussion.

14 Apr

[Lead: G4] Ghodsi et al. Dominant resource fairness: Fair allocation of multiple resource types, NSDI 2011.

12 Apr

[Lead: G3] Qureshi et al. Cutting the electrical bill for Internet-scale systems, SIGCOMM 2009.

7 Apr

[Lead: G2] Meisner et al. Power management of on-line data intensive services, ISCA 2011.

31 Mar

[Lead: G1] Meisner et al. PowerNap: Eliminating server idle power, ASPLOS 2009.

29 Mar

[Lead: G5] Fan et al. Power provisioning for a warehouse-sized computer, ISCA 2007.

24 Mar

[Lead: G4] Abts et al. Energy proportional datacenter networks, ISCA 2010.

22 Mar

[Lead: G3] Heller et al. ElasticTree: Saving energy in data center networks, NSDI 2010.

10 Mar

[Lead: G2] Ousterhout et al. The Case for RAMCloud, CACM, 2011

3 Mar

[Lead: G1] Novakovic et al. Scale-out NUMA, ASPLOS 2014.

25 Feb

[Lead: G5] Lim et al. Thin servers with smart pipes: Designing SoC accelerators for Memcached, ISCA 2013.

23 Feb

[Lead: G4] Putnam et al., A reconfigurable fabric for accelerating large-scale datacenter services, ISCA 2014.

18 Feb

[Lead: G3] Lo et al., Towards energy proportionality for large-scale latency-critical workloads, ISCA 2014.

16 Feb

[Lead: G2] Reddi et al., Web search using mobile cores: Quantifying and mitigating the price of efficiency, ISCA 2010.

11 Feb

[Lead: G1] Dean and Barroso. The tail at scale, CACM 2013.

9 Feb

[Lead: G5] Lim et al. MICA: A holistic approach to fast in-memory key-value storage, NSDI 2014.

4 Feb

[Lead: G4] Zaharia et al. Resilient distributed datasets: A fault-tolerant abstraction for in-memory cluster computing, NSDI 2012.

2 Feb

[Lead: G3] Dean and Ghemawat. MapReduce: Simplified data processing on large clusters, OSDI, 2004.

28 Jan

[Lead: G2] Ghemawat et al. The Google File System, SOSP 2003.

26 Jan

[Lead: G1] Barroso, Cildaras, Hoelzle. The datacenter as a computer: An introduction to the design of warehouse-scale machines," Synthesis Lectures on Computer Architecture, 2013. (Chapters 5-8).

21 Jan

No class.

19 Jan

[Lead: BL] Barroso, Cildaras, Hoelzle. The datacenter as a computer: An introduction to the design of warehouse-scale machines," Synthesis Lectures on Computer Architecture, 2013. (Chapters 1-4)

14 Jan

Introduction / Open Discussion