Hello! I'm a third year PhD Student in the Reliable Adaptive Distributed System Lab at UC Berkeley. My interests are in scalable structured storage for large interactive web applications.
Projects
SCADS
Collaborative web applications such as Facebook, Flickr and Yelp present new challenges for storing and querying large amounts of data. As users and developers are focused more on performance than single copy consistency or the ability to perform ad-hoc queries, there exists an opportunity for a highly-scalable system tailored specifically for relaxed consistency and pre-computed queries. The Web 2.0 development model demands the ability to both rapidly deploy new features and automatically scale with the number of users. There have been many successful distributed key-value stores, but so far none provide as rich a query language as SQL. We propose a new architecture, SCADS, that allows the developer to declaratively state application specific consistency requirements, takes advantage of utility computing to provide cost effective scale-up and scale-down, and will use machine learning models to introspectively anticipate performance problems and predict the resource requirements of new queries before execution.
- SCADS: Scale-Independent Storage for Social Computing Applications (CIDR09)
- Spring 2008 Retreat Talk
- Retreat Poster
Cloud Computing
In addition to being an active user of cloud computing for my own research (mostly Amazon's EC2), I was also co-author on a technical report about cloud computing along with a number of professors and students in the RAD Lab.
- Above the Clouds: A Berkeley View of Cloud Computing Blog
- Above the Clouds Technical Report
- Above the Clouds Presentation (pdf)
The Declarative Datacenter
As datacenters grow dramatically in size, they are becoming increasingly difficult to operate. Our vision is to create a declarative language to specify what services will run in a datacenter. This specification will drive a new datacenter operating system, which will dynamically allocate resources based on changing application needs.
Ruby On Rails
A lot of people in the RAD lab think Rails is pretty cool, so I occasinally help out with a class on Next-Generation Internet Services With Ruby on Rails. I also spent some free time working with other students on an overly complicated Awesome Movie Voting System for my labs weekly social movie night.
Drape
Dynamic Resource Allocation for Power Efficiency investigates methods for conserving power in large data centers. Our initial prototype was a system that used a predicted workload to turn servers on and off while replaying traces from the france98 worldcup website. The DRAPE project report for CS262a is available online.
Teaching
Fall 2008 - CS186 - Introduction to Database Management Systems along with Michael Franklin and Beth Trushkowsky.
Contact Info
University of California at Berkeley
RAD Lab - MS 1776
465 Soda Hall, Room 477
Berkeley CA 94720
email: marmbrus at cs.berkeley.edu
Random Michael Links
Here is a random collection of other places you can find me on the internet. It's mostly an experiment with page-rank to see if I can take over the whole first page when you google my name.