@device(postscript)
@libraryfile(Mathematics10)
@libraryfile(Accents)
@style(fontfamily=timesroman,fontscale=11)
@pagefooting(immediate, left "@c<Technical Report 1996>",
center "@c<Computer Science>",
right "@c<Carnegie Mellon>")
@heading(Application-Allocated I/O Buffering with
System-Allocated Performance)
@heading(CMU-CS-96-169)
@center(@b(Jose Carlos Brustoloni, Peter Steenkiste))
@center(August 1996@foot<An abridged version of this paper appears as "Effects of Buffering Semantics on I/O Performance" in @i[Proceedings of the Second Symposium on Operating Systems Design and Implementation, OSDI'96], USENIX, Seattle, Washignton, October 28-31, 1996, pp. 277-291, and in @i[Operating Systems Review], SIGOPS, ACM, vol. 30, Special Issues, 1996, pp. 277-291.>)
@center(FTP: CMU-CS-96-169.ps)
@blankspace(1)
@begin(text)

We present a novel taxonomy that classifies I/O data passing
between applications and operating system along three dimensions:
buffer allocation scheme, guaranteed integrity, and optimization.
We contribute new optimizations @y(M) @i(input-disabled pageout, 
transient output copy-on-write), and @i(input alignment) @y(M) that are 
used in a novel buffering semantics, @i(emulated copy).
We implemented an I/O framework, Genie, that allows applications to
select any semantics in the taxonomy. Using Genie for end-to-end 
communication over an ATM network, we found that, compared to 
other semantics, only copy had sharply inferior performance. 
All other semantics performed quite similarly, contradicting the 
expectation that emulated copy, being application-allocated and 
strong-integrity (as is copy semantics), should have considerably 
worse performance than those of move (system-allocated) or share 
(weak-integrity) semantics. We analyzed end-to-end latencies in 
terms of the costs of primitive data passing operations and modeled 
how those costs scale with CPU, memory, and network speeds. 
The analysis suggests that current trends tend to intensify the 
observed performance clustering. We conclude that I/O interfaces 
with copy semantics, such as that of Unix, can be transparently 
converted to emulated copy and thus achieve performance approaching 
the best in the taxonomy.

@blankspace(2line)
@begin(transparent,size=10)
@b(Keywords:@ )@c<Buffering, copy avoidance, I/O, data communications,
performance analysis, operating systems, ATM networks>
@end(transparent)
@blankspace(1line)
@end(text)
@flushright(@b[(29 pages)])