FFS (Fast Flexible Serialization) is a middleware library for data
representation, processing and marshaling that preserves the performance
of traditional approaches while relaxing the requirement of a priori knowledge and providing
complex run-time flexibility. FFS provides for highly efficient binary data
communication, XML-like examination of unknown data, and both third-party
and in situ data processing via dynamic code generation. All of these
capabilities are fully dynamic at run-time, without requiring a priori
agreements or knowledge of the exact form of the data being communicated or
analyzed.
The efficient binary representation of the data object is also
realized in the output format for the FFSfile, a file format
that supports attribute based querying for data objects while providing transparent
support for machine portability and in situ processing.
FFS has been used in a number of Georgia Tech systems and projects. The
manual for FFS is not yet finalized, but is
available
for review, and a paper describing its general principles and operation is here:
A Type System for High Performance Communication and Computation,
Greg Eisenhauer, Matthew Wolf, Hasan Abbasi, Scott Klasky, Karsten Schwan.
to appear at the Workshop on D3Science associated with e-Science 11, December 5, 2011, Stockholm, Sweden
PDF Format [160K]
Getting FFS.
FFS is available under the new BSD license.
A source distribution is available in gzip format.
Alternatively, for binary installations we have :
This page is maintained by
Greg Eisenhauer