Coming Soon in 6.853 - Week of 5/9

This week's hand-outs include four sets of SOSP submissions. Take only the set you have been assigned; if you wish to review additional papers, we will arrange that later. Use of the submissions has been approved by the program committee and authors with two conditions attached:

• Please retain the student comments, and provide a way to get them to me (Mark Weiser) so I can get them to the authors. The authors consider this their quid pro quo, that they get more feedback.
• Please inform your students that they should not carry on additional work based on their reading of these papers without first contacting the authors or waiting until actual publication in December. This is the usual professional courtesy that one does not act on work one comes across as a reviewer without contacting the authors. You can find out all the authors after the PC meeting in June.

Virtual Memory & Filesystems (filesystems@tns.lcs.mit.edu)

• Performance of Cache Coherence in Stackable Filing #56
• Serverless Network File Systems #77
• Logged Virtual Memory: System Support for Simulation, Transactions, and Debugging #79

Network Objects (objects@tns.lcs.mit.edu)

• Using Dynamic Sets to Speed Search in World Wide Information Systems #65
• Secure Network Objects #55
• A Hierarchical Internet Object Cache #27
• Object and Native Code Process Mobility Among Hetrogeneous Computers #52

Kernels (kernels@tns.lcs.mit.edu)

• On Micro-Kernel Construction #49
• A New Page Table for 64-bit Address Space #66
• Specializing Object-Oriented RPC for Functionality and Performance #80

Grab Bag (grabs@tns.lcs.mit.edu)

• Hive: Fault Containment for Shared-Memory Multiprocessors #78
• The AutoRAID hierarchical storage system #20
• Exploiting Process Lifetime Distributions for Dynamic Load Balancing #14
• U-Net: A User-Level Network Interface for Parallel and Distributed Computing #13

Your Task

Your task is to review your set of papers, and to each submit one filled-in review form per paper to the appropriate mailing list. We also hope that you will use the mailing lists to function as a program committee (figure out the organization yourselves) and come up with an overall ranking of your papers. We will call on members of the program committees to present their findings during class. The reviews are part of your class participation, and will not be otherwise graded.