The first assignment, done individually or in pairs, is to try to quantify
one of the potential benefits of Intelligent Disks or Intelligent
Main Memory. A list of initial experiment suggestions will appear on this
page. The paper and presentation should have some quantitative evaluation.
Pointers to results from Assignment 1:
vs. Linux : compare performance of an appliance (Snap! NFS Server)
to a standard PC (Linux NFS Server) using the xFS microbenchmarks and the
Bakeoff for ISTORE : Comparing Linix, QNX, NT for OS benchmarks
Appliance Survey : A survey of existing (or former) products
loosely categorized as "network appliances".
Research in the Berkeley EECS Department : A survey of research projects
which are forming the foundation of the post-pc era
Index of /~randit/slides
: Investigation of errors recorded from 6 months of Tertiary Disk operation
of 395 disks, 20 PCs
IDISK : An
alternative design to IRAM with the the disk controllers on the same
die as the processor and memory plus a new organization of the disk drive
Researchers have been criticized recently for not thinking about big
research ideas, putting their projects being barely ahead of commercial
developments. Perhaps this is because we're not asked to do things like
this in grad school?
This second assignment, done individually or in pairs, is to Think
Our focus is on the next computer generation, the so-called "Post-PC
Era". Discuss some of the big issues that must be addressed as part of
the next revolution: benchmarks, legacy code, requirements for new businesses
in this era, ...
This page last updated on November 12, 1998.