Generate redundant blocks of information such that if some of the blocks are
lost then the original data can be recovered from the remaining blocks. This
package includes command-line tools, C API, Python API, and Haskell API.
Intro and Licence
This package implements an “erasure code”, or “forward error correction
You may use this package under the GNU General Public License, version 2 or,
at your option, any later version. You may use this package under the
Transitive Grace Period Public Licence, version 1.0 or, at your option, any
later version. (You may choose to use this package under the terms of either
licence, at your option.) See the file COPYING.GPL for the terms of the GNU
General Public License, version 2. See the file COPYING.TGPPL.rst for the
terms of the Transitive Grace Period Public Licence, version 1.0.
The most widely known example of an erasure code is the RAID-5 algorithm
which makes it so that in the event of the loss of any one hard drive, the
stored data can be completely recovered. The algorithm in the zfec package
has a similar effect, but instead of recovering from the loss of only a
single element, it can be parameterized to choose in advance the number of
elements whose loss it can tolerate.
This package is largely based on the old “fec” library by Luigi Rizzo et al.,
which is a mature and optimized implementation of erasure coding. The zfec
package makes several changes from the original “fec” package, including
addition of the Python API, refactoring of the C API to support zero-copy
operation, a few clean-ups and optimizations of the core code itself, and the
addition of a command-line tool named “zfec”.
pip install zfec
To run the self-tests, execute tox from an unpacked source tree or git checkout.
To run the tests of the Haskell API: runhaskell haskell/test/FECTest.hs
Note that in order to run the Haskell API tests you must have installed the
library first due to the fact that the interpreter cannot process FEC.hs as
it takes a reference to an FFI function.
This package performs two operations, encoding and decoding. Encoding takes
some input data and expands its size by producing extra “check blocks”, also
called “secondary blocks”. Decoding takes some data – any combination of
blocks of the original data (called “primary blocks”) and “secondary blocks”,
and produces the original data.
The encoding is parameterized by two integers, k and m. m is the total
number of blocks produced, and k is how many of those blocks are necessary to
reconstruct the original data. m is required to be at least 1 and at most
256, and k is required to be at least 1 and at most m.
(Note that when k == m then there is no point in doing erasure coding – it
degenerates to the equivalent of the Unix “split” utility which simply splits
the input into successive segments. Similarly, when k == 1 it degenerates to
the equivalent of the unix “cp” utility – each block is a complete copy of
the input data.)
Note that each “primary block” is a segment of the original data, so its size
is 1/k’th of the size of original data, and each “secondary block” is of the
same size, so the total space used by all the blocks is m/k times the size of
the original data (plus some padding to fill out the last primary block to be
the same size as all the others). In addition to the data contained in the
blocks themselves there are also a few pieces of metadata which are necessary
for later reconstruction. Those pieces are: 1. the value of K, 2. the
value of M, 3. the sharenum of each block, 4. the number of bytes of
padding that were used. The “zfec” command-line tool compresses these pieces
of data and prepends them to the beginning of each share, so each the
sharefile produced by the “zfec” command-line tool is between one and four
bytes larger than the share data alone.
The decoding step requires as input k of the blocks which were produced by
the encoding step. The decoding step produces as output the data that was
earlier input to the encoding step.
To run the benchmarks, execute the included bench/bench_zfec.py script with
optional –k= and –m= arguments.
On my Athlon 64 2.4 GHz workstation (running Linux), the “zfec” command-line
tool encoded a 160 MB file with m=100, k=94 (about 6% redundancy) in 3.9
seconds, where the “par2” tool encoded the file with about 6% redundancy in
27 seconds. zfec encoded the same file with m=12, k=6 (100% redundancy) in
4.1 seconds, where par2 encoded it with about 100% redundancy in 7 minutes
and 56 seconds.
The underlying C library in benchmark mode encoded from a file at about 4.9
million bytes per second and decoded at about 5.8 million bytes per second.
On Peter’s fancy Intel Mac laptop (2.16 GHz Core Duo), it encoded from a file
at about 6.2 million bytes per second.
On my even fancier Intel Mac laptop (2.33 GHz Core Duo), it encoded from a
file at about 6.8 million bytes per second.
On my old PowerPC G4 867 MHz Mac laptop, it encoded from a file at about 1.3
million bytes per second.
Here is a paper analyzing the performance of various erasure codes and their
implementations, including zfec:
Zfec shows good performance on different machines and with different values
of K and M. It also has a nice small memory footprint.
Each block is associated with “blocknum”. The blocknum of each primary block
is its index (starting from zero), so the 0’th block is the first primary
block, which is the first few bytes of the file, the 1’st block is the next
primary block, which is the next few bytes of the file, and so on. The last
primary block has blocknum k-1. The blocknum of each secondary block is an
arbitrary integer between k and 255 inclusive. (When using the Python API,
if you don’t specify which secondary blocks you want when invoking encode(),
then it will by default provide the blocks with ids from k to m-1 inclusive.)
fec_encode() takes as input an array of k pointers, where each pointer
points to a memory buffer containing the input data (i.e., the i’th buffer
contains the i’th primary block). There is also a second parameter which
is an array of the blocknums of the secondary blocks which are to be
produced. (Each element in that array is required to be the blocknum of a
secondary block, i.e. it is required to be >= k and < m.)
The output from fec_encode() is the requested set of secondary blocks which
are written into output buffers provided by the caller.
Note that this fec_encode() is a “low-level” API in that it requires the
input data to be provided in a set of memory buffers of exactly the right
sizes. If you are starting instead with a single buffer containing all of
the data then please see easyfec.py’s “class Encoder” as an example of how
to split a single large buffer into the appropriate set of input buffers
for fec_encode(). If you are starting with a file on disk, then please see
filefec.py’s encode_file_stringy_easyfec() for an example of how to read
the data from a file and pass it to “class Encoder”. The Python interface
provides these higher-level operations, as does the Haskell interface. If
you implement functions to do these higher-level tasks in other languages,
please send a patch to firstname.lastname@example.org so that your API can be
included in future releases of zfec.
fec_decode() takes as input an array of k pointers, where each pointer
points to a buffer containing a block. There is also a separate input
parameter which is an array of blocknums, indicating the blocknum of each
of the blocks which is being passed in.
The output from fec_decode() is the set of primary blocks which were
missing from the input and had to be reconstructed. These reconstructed
blocks are written into output buffers provided by the caller.
encode() and decode() take as input a sequence of k buffers, where a
“sequence” is any object that implements the Python sequence protocol (such
as a list or tuple) and a “buffer” is any object that implements the Python
buffer protocol (such as a string or array). The contents that are
required to be present in these buffers are the same as for the C API.
encode() also takes a list of desired blocknums. Unlike the C API, the
Python API accepts blocknums of primary blocks as well as secondary blocks
in its list of desired blocknums. encode() returns a list of buffer
objects which contain the blocks requested. For each requested block which
is a primary block, the resulting list contains a reference to the
apppropriate primary block from the input list. For each requested block
which is a secondary block, the list contains a newly created string object
containing that block.
decode() also takes a list of integers indicating the blocknums of the
blocks being passed int. decode() returns a list of buffer objects which
contain all of the primary blocks of the original data (in order). For
each primary block which was present in the input list, then the result
list simply contains a reference to the object that was passed in the input
list. For each primary block which was not present in the input, the
result list contains a newly created string object containing that primary
Beware of a “gotcha” that can result from the combination of mutable data
and the fact that the Python API returns references to inputs when
Returning references to its inputs is efficient since it avoids making an
unnecessary copy of the data, but if the object which was passed as input
is mutable and if that object is mutated after the call to zfec returns,
then the result from zfec – which is just a reference to that same object
– will also be mutated. This subtlety is the price you pay for avoiding
data copying. If you don’t want to have to worry about this then you can
simply use immutable objects (e.g. Python strings) to hold the data that
you pass to zfec.
The Haskell code is fully Haddocked, to generate the documentation, run
runhaskell Setup.lhs haddock.
The filefec.py module has a utility function for efficiently reading a file
and encoding it piece by piece. This module is used by the “zfec” and
“zunfec” command-line tools from the bin/ directory.
A C compiler is required. To use the Python API or the command-line tools a
Python interpreter is also required. We have tested it with Python v2.7,
v3.5 and v3.6. For the Haskell interface, GHC >= 6.8.1 is required.
Thanks to the author of the original fec lib, Luigi Rizzo, and the folks that
contributed to it: Phil Karn, Robert Morelos-Zaragoza, Hari Thirumoorthy, and
Dan Rubenstein. Thanks to the Mnet hackers who wrote an earlier Python
wrapper, especially Myers Carpenter and Hauke Johannknecht. Thanks to Brian
Warner and Amber O’Whielacronx for help with the API, documentation,
debugging, compression, and unit tests. Thanks to Adam Langley for improving
the C API and contributing the Haskell API. Thanks to the creators of GCC
(starting with Richard M. Stallman) and Valgrind (starting with Julian
Seward) for a pair of excellent tools. Thanks to my coworkers at Allmydata
– http://allmydata.com – Fabrice Grinda, Peter Secor, Rob Kinninmont, Brian
Warner, Zandr Milewski, Justin Boreta, Mark Meras for sponsoring this work
and releasing it under a Free Software licence. Thanks to Jack Lloyd, Samuel
Neves, and David-Sarah Hopwood.