otdrs
is a Rust implementation of a SOR file parser and generator. SOR files are used as a storage format for optical time-domain reflectometry (OTDR) tests, which are commonly used to characterise and validate the proper construction of fibre optic networks. SOR is formally known as the Bellcore OTDR interchange format, specified in Telcordia SR-4731, a proprietary standard. OTDR testing involves firing short pulses of light down fibres under test, and measuring the intensity of returned light over time.
otdrs
is intended as a minimal but valid and robust parser implementation to enable a translation from the closed binary format to open, self-documenting formats, to permit easy development of analysis tools without having to deal with the complexities of managing a binary format that is broadly undocumented and difficult to parse. It also provides Rust primitives for OTDR files and a writer to allow files to be loaded, modified, and written back.
Wherever possible, translation preserves the input format; otdrs does not try to fix peculiarities of particular test equipment or post-processing software or work around quirks in their input, except where they would break parsing of the format. No normalisation is attempted.
Where the parser cannot reliably extract information it is omitted; in this sense, otdrs
is a best-effort parser.
Writing of files is performed with as much care as possible and the optional checksum block is computed. Generated files open without issue in several professional applications, and proprietary data is preserved losslessly.
Rust was chosen for its robustness, type-safety, and the excellent nom
parser library. serde
is used for serialisation for output.
otdrs
takes one positional argument, the path to a SOR file. Its output is a single JSON or CBOR blob which contains the information within the SOR file; flags are used to set the output path (default is stdout) or the format to output. otdrs --help
shows the available options.
A post-processing example is shown in the demo.py
script in this repository, which will plot the data from an OTDR file.
If you have Rust/Cargo installed you can install otdrs with cargo install otdrs
. otdrs
is not otherwise packaged currently.
You can use otdrs as a Rust library to read, modify, and write SOR files. For instance, the following minimal program will modify the fiber ID and cable ID of the file, and print the nominal wavelength of the file to the terminal.
use otdrs::parser::parse_file;
use std::fs::File;
use std::io::prelude::*;
fn main() -> std::io::Result<()> {
// Read the file into memory
let mut in_file = File::open("input.sor")?;
let mut in_data = Vec::new();
in_file.read_to_end(&mut in_data)?;
// Parse the file
let mut sor = parse_file(&in_data).unwrap().1;
// Most blocks are Options because the parser can't guarantee them
// This is true even of "required" blocks in the spec, because otdrs is permissive.
match sor.general_parameters.as_mut() {
Some(mut gp) => {
// If we've got a block we can read and modify it!
println!("Nominal wavelength of this SOR record: {:?}", gp.nominal_wavelength);
gp.fiber_id = "Hello world";
gp.cable_id = "Foo bar";
// Note that we don't check the values you write a valid, outside of type enforcement
// Even then you need to be careful to adhere to the spec. e.g. no UTF-8 glyphs
// No emojis in your comments field ๐ญ๐ญ๐ญ
}
None => {
println!("Your SOR file has no General Parameters block!")
}
}
// Encode the SORFile structure to a binary SOR file again...
// Note you'd normally want to handle errors properly here, but we're an example, so...
let bytes_to_write = sor.to_bytes().unwrap();
// Write the file out to disk!
let mut out_file = File::create("output.sor").unwrap();
out_file.write_all(&bytes_to_write)?;
return Ok(());
}
This is the author's first major Rust project, so use with caution.
This is absolutely not guaranteed to work in all cases with all files or produce correct output in all cases (or indeed any). Check the output with the files you're using and a known-good viewer if you care about what you're doing.
As otdrs
handles arbitrary binary input and performs some arithmetic on it which can potentially lead to underruns and overruns as well as exciting undefined behaviour.
While Rust is a very good language to write such tools in, since runtime errors such as this are handled, otdrs
makes an effort to avoid obvious situations where slice pointers violate bounds or where arithmetic on SOR contents may lead to unexpected situations. cargo-fuzz is used to fuzz with libFuzzer (on Linux only at present) to discover scenarios in which this can occur, including through writing and round-tripping OTDR files.
Checking of result validity is not performed on all fields, and users of the tool should take care to avoid trusting input parsed from SOR files. Sanitise your inputs.
- The "link parameters" block is not currently decoded, as the author does not have files which contain it for testing. This is not used in common OTDR sets.
- Testing is not as comprehensive and extensive as it should be, particularly for writing files.
There is no application of fixed scaling factors described in SR-4731. This is generally intentional, to permit correct post-processing as required in other applications.
While SOR files are standardised, not all of the content is; there are a set of standard and required blocks for the basic information, and otdrs only attempts to parse the standard blocks in a SOR file.
The content of proprietary blocks is dumped for analysis by upstream tools that may either have knowledge of proprietary formats or wish to simply know of the existence of such blocks. The map block will in all cases list all blocks within the file.
otdrs
has experimental support for generating SORs from Rust data structures. Strictly, the map block is heavily recomputed when writing; a BlockInfo block with a revision number and header will be expected for all blocks, but sizes and counters are dynamically generated. This is because it is practically impossible (or very difficult, at least) to compute sizes before serialising data, so this is best done at the point of writing.
Editors are responsible for ensuring that any modification of data elsewhere in the file makes sense, e.g. if the number of points within a DataPointsAtScaleFactor
struct is changed, then the n_points
field must be amended by the editor; otdrs
will not do this for you.
Currently, landmarks and LinkParameters
are not written out, as these are very rarely used in practice and no example data is currently available to support testing.
The parser has been tested on SOR files generated from:
- Noyes OFL280 OTDRs, including those re-exported from EXFO FastReporter3
- Anritsu Access Master OTDRs
- EXFO MaxTester 730C and FTB-4/FTBx730 OTDRs
- EXFO iOLM files exported to SOR from EXFO FastReporter3
- EXFO 735C-SM7R modules in centralised PON test deployments (shooting through splitters to high-reflectance devices)
Round-trip testing has also confirmed that all of the above files can be read and written out by otdrs
with no loss or alteration of data.
Further test files are desired and should be submitted to the author or as a pull request with tests against known values.
To actually interpet any of this data correctly you are probably going to need to read SR-4731, which can be found here for around $750.
This parser makes no attempt to correctly interpret the resulting data from the SOR file format, merely to make it accessible for applications to perform correct interpretation. Actually locating events and measuring cable data based on OTDR data requires careful consideration of data offsets (e.g. front panel to user offset, scaling factors, etc).
As with pretty much every standard out there, every vendor has interpreted it differently. For one example, some Noyes OTDRs store a 30 second averaging period as 3000
whereas EXFO and Anritsu record it as 30
. The specification, of course, says this should be stored as 300
. Professional software struggles with this - for instance, EXFO's FastReporter3 software misinterprets the Noyes result.
Documentation of storage quirks compared to "standard" behaviour such as the above, against known behaviour would be helpful for those developing post-processing software; if you have access to test equipment, you can helpfully run a controlled test, write down the actual values displayed by the tester etc and store the SOR file directly.
While "fixing" vendor quirks to generate a standards-compliant output is not currently in the scope of otdrs
, this is something that could be added as an optional post-processing step.
- 1.0.0 - refactored to avoid some beginner Rust errors; SORFile now owns its data. Updated dependencies.
- 0.4.2 - upgraded nom to 7.1.0, clap to 3.0.0-rc7
- 0.4.1 - upgraded nom to 6.1.2, improved README and demo scripts
- 0.4.0 - added SORFile#to_bytes and a whole bunch of related functions and macros which altogether mean that
otdrs
can now write OTDR files - 0.3.0 - switched to using clap for command-line argument handling for better error handling, added option to write to file instead of stdout and added CBOR export support
- 0.2.0 - restructured to allow use as a library, added fuzzing and fixed a number of bounds-check/error propogation problems
- 0.1.0 - initial release
GPLv3 has been selected specifically to drive improved open source engagement with equipment manufacturers and developers of OTDR processing software in an industry that has struggled with open data exchange, proprietary (and vendor-locked) software, and poor maintenance of existing software.
otdrs - a SOR file parsing tool Copyright (C) 2021 James Harrison
This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.