EVTX

This is a parser for the Windows EVTX format.

Note that it is complete as in the sense that it successfully parses a wide variety of samples, but I've yet to implement the full specification.

This parser is implemented using 100% safe rust, and should work on recent (i'm testing against 1.34) versions of rust.

Documentation

Python bindings are available as well at https://github.com/omerbenamram/pyevtx-rs (still experimental, will publish to PyPi soon)

Example usage (associated binary utility):

• cargo install evtx
• run evtx_dump --input <evtx_file> to dump contents of evtx records as xml.

Example usage (as library):

    use evtx::EvtxParser;
    
    fn main() {
        let parser = EvtxParser::from_path(fp).unwrap();
        for record in parser.records() {
            match record {
                Ok(r) => println!("Record {}\n{}", r.event_record_id, r.data),
                Err(e) => eprintln!("{}", e),
            }
        }
    }

For parallel iteration (uses rayon):

    use evtx::EvtxParser;
    
    fn main() {
        let parser = EvtxParser::from_path(fp).unwrap();
        for record in parser.parallel_records() {
            match record {
                Ok(r) => println!("Record {}\n{}", r.event_record_id, r.data),
                Err(e) => eprintln!("{}", e),
            }
        }
    }

The parallel version is enabled when compiling with feature "multithreading" (enabled by default).

Benchmarking

Initial benchmarking I've performed indicate that this implementation is probably the fastest available 🍺.

I'm using a real world, 30MB sample which contains ~62K records.

This is benchmarked on my 2017 MBP.

Comparison with other libraries:

• python-evtx (https://github.com/williballenthin/python-evtx)

  With CPython this is quite slow

  time -- python3 ~/Workspace/python-evtx/scripts/evtx_dump.py ./samples/security_big_sample.evtx > /dev/null                                                                      Mon Apr  1 19:41:16 2019
        363.83 real       356.26 user         2.17 sys
  

  With PyPy (tested with pypy3.5, 7.0.0), it's taking just less than a minute (a 6x improvement!)

  time -- pypy3 ~/Workspace/python-evtx/scripts/evtx_dump.py ./samples/security_big_sample.evtx > /dev/null                                                                      Mon Apr  1 19:41:16 2019
        59.30 real        58.10 user         0.51 sys
  

• libevtx (https://github.com/libyal/libevtx)

  This library is written in C, so I initially expected it to be faster than my implementation.

  It clocks in about 6x faster than PyPy.

  time -- ~/Workspace/libevtx/dist/bin/evtxexport -f xml ./samples/security_big_sample.evtx > /dev/null
         11.30 real        10.77 user         0.41 sys
  

  Note: libevtx does have multi-threading support planned (according to the readme), but isn't implemented as of writing this (April 2019).

• evtx (this library!)

  When using a single thread, this implementation is about 2x faster than C

  time -- ./target/release/main --input ./samples/security_big_sample.evtx > /dev/null                                                                                     516ms  Mon Apr  1 19:53:59 2019
          4.65 real         4.53 user         0.10 sys
  

  With multi-threading enabled, it blazes through the file in just 1.5 seconds:

  time -- ./target/release/main -t --input ./samples/security_big_sample.evtx > /dev/null                                                                                 4661ms  Mon Apr  1 19:54:14 2019
          1.51 real         7.50 user         0.26 sys
  

Caveats

• I haven't implemented any sort of recovery/carving of records (available in some other implementations).
• I haven't tested this against samples which contains esotericlly encoded strings.

License

Licensed under either of

• Apache License, Version 2.0, (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
• MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)

at your option.

Contribution

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.