220 likes | 339 Views
Predictable Computer Systems. Remzi Arpaci-Dusseau University of Wisconsin, Madison. Trends. Trends. Complexity Cheap Components Everything Interconnected. Problems. Nothing Works As Expected. Performance Fault-Tolerance Security. What Would Be Ideal. Ideal.
E N D
Predictable Computer Systems Remzi Arpaci-Dusseau University of Wisconsin, Madison
Trends • Complexity • Cheap Components • Everything Interconnected
Nothing Works As Expected • Performance • Fault-Tolerance • Security
Ideal • Assemble large-scale systemfrom cheap, complex components • System works in predictable manner
State of the Art • APIs • Protocols
A Small Example:Understanding the Failure Behaviorof Local File Systems
Understanding FS Failure • Type-aware fault injection • Make fault injection layer aware of FS structures • e.g., make an inode block fail • Why useful • Can infer how file system reacts to failuresat different points in its code
Ext3 ReiserFS JFS Zero Stop Propagate Recovery Retry Redundancy Write Errors: Recovery Techniques • Ext3, JFS don’treact to write failures • ReiserFS (almost) always calls panic()
What We Need • Vocabulary • + • Techniques • + • Tools • Methods to • = Understand • Behavior • Predictable • -> Computer • Systems
ADvanced Systems Lab (ADSL) • Gray-box Operating Systems and Storage Systems Andrea Arpaci-Dusseau Remzi Arpaci-Dusseau
ADvanced Systems Lab (ADSL) • Who does the real work: • Nitin Agrawal • Lakshmi Bairavasundaram • John Bent • Nathan Burnett • Tim Denehy • Camille Fournier • Haryadi Gunawi • Todd Jones • James Nugent • Ina Popovici • Vijayan Prabhakaran • Muthian Sivathanu • Who does the real work: • Nitin Agrawal • Lakshmi Bairavasundaram • John Bent • Nathan Burnett • Tim Denehy • Camille Fournier • Haryadi Gunawi • Todd Jones • James Nugent • Ina Popovici • Vijayan Prabhakaran • Muthian Sivathanu
Large-Scale Distributed Systems Front Ends C DBMS W D C D W Internet Net W C Clients Archival Storage Online Storage
Ideal: Legos Top Side What You See Is What You Get