encoding

Go package containing implementations of encoders and decoders for various data formats.

Motivation

At Segment, we do a lot of marshaling and unmarshaling of data when sending, queuing, or storing messages. The resources we need to provision on the infrastructure are directly related to the type and amount of data that we are processing. At the scale we operate at, the tools we choose to build programs can have a large impact on the efficiency of our systems. It is important to explore alternative approaches when we reach the limits of the code we use.

This repository includes experiments for Go packages for marshaling and unmarshaling data in various formats. While the focus is on providing a high performance library, we also aim for very low development and maintenance overhead by implementing APIs that can be used as drop-in replacements for the default solutions.

Requirements and Maintenance Schedule

This package has no dependencies outside of the core runtime of Go. It requires a recent version of Go.

This package follows the same maintenance schedule as the Go project, meaning that issues relating to versions of Go which aren't supported by the Go team, or versions of this package which are older than 1 year, are unlikely to be considered.

Additionally, we have fuzz tests which aren't a runtime required dependency but will be pulled in when running go mod tidy. Please don't include these go.mod updates in change requests.

encoding/json

More details about how this package achieves a lower CPU and memory footprint can be found in the package README.

The json sub-package provides a re-implementation of the functionalities offered by the standard library's encoding/json package, with a focus on lowering the CPU and memory footprint of the code.

The exported API of this package mirrors the standard library's encoding/json package, the only change needed to take advantage of the performance improvements is the import path of the json package, from:

import (
    "encoding/json"
)

to

import (
    "github.com/segmentio/encoding/json"
)

The improvement can be significant for code that heavily relies on serializing and deserializing JSON payloads. The CI pipeline runs benchmarks to compare the performance of the package with the standard library and other popular alternatives; here's an overview of the results:

Comparing to encoding/json (v1.16.2)

name                           old time/op    new time/op     delta
Marshal/*json.codeResponse2      6.40ms ± 2%     3.82ms ± 1%   -40.29%  (p=0.008 n=5+5)
Unmarshal/*json.codeResponse2    28.1ms ± 3%      5.6ms ± 3%   -80.21%  (p=0.008 n=5+5)

name                           old speed      new speed       delta
Marshal/*json.codeResponse2     303MB/s ± 2%    507MB/s ± 1%   +67.47%  (p=0.008 n=5+5)
Unmarshal/*json.codeResponse2  69.2MB/s ± 3%  349.6MB/s ± 3%  +405.42%  (p=0.008 n=5+5)

name                           old alloc/op   new alloc/op    delta
Marshal/*json.codeResponse2       0.00B           0.00B           ~     (all equal)
Unmarshal/*json.codeResponse2    1.80MB ± 1%     0.02MB ± 0%   -99.14%  (p=0.016 n=5+4)

name                           old allocs/op  new allocs/op   delta
Marshal/*json.codeResponse2        0.00            0.00           ~     (all equal)
Unmarshal/*json.codeResponse2     76.6k ± 0%       0.1k ± 3%   -99.92%  (p=0.008 n=5+5)

Benchmarks were run on a Core i9-8950HK CPU @ 2.90GHz.

Comparing to github.com/json-iterator/go (v1.1.10)

name                           old time/op    new time/op    delta
Marshal/*json.codeResponse2      6.19ms ± 3%    3.82ms ± 1%   -38.26%  (p=0.008 n=5+5)
Unmarshal/*json.codeResponse2    8.52ms ± 3%    5.55ms ± 3%   -34.84%  (p=0.008 n=5+5)

name                           old speed      new speed      delta
Marshal/*json.codeResponse2     313MB/s ± 3%   507MB/s ± 1%   +61.91%  (p=0.008 n=5+5)
Unmarshal/*json.codeResponse2   228MB/s ± 3%   350MB/s ± 3%   +53.50%  (p=0.008 n=5+5)

name                           old alloc/op   new alloc/op   delta
Marshal/*json.codeResponse2       8.00B ± 0%     0.00B       -100.00%  (p=0.008 n=5+5)
Unmarshal/*json.codeResponse2    1.05MB ± 0%    0.02MB ± 0%   -98.53%  (p=0.000 n=5+4)

name                           old allocs/op  new allocs/op  delta
Marshal/*json.codeResponse2        1.00 ± 0%      0.00       -100.00%  (p=0.008 n=5+5)
Unmarshal/*json.codeResponse2     37.2k ± 0%      0.1k ± 3%   -99.83%  (p=0.008 n=5+5)

Although this package aims to be a drop-in replacement of encoding/json, it does not guarantee the same error messages. It will error in the same cases as the standard library, but the exact error message may be different.

encoding/iso8601

The iso8601 sub-package exposes APIs to efficiently deal with with string representations of iso8601 dates.

Data formats like JSON have no syntaxes to represent dates, they are usually serialized and represented as a string value. In our experience, we often have to check whether a string value looks like a date, and either construct a time.Time by parsing it or simply treat it as a string. This check can be done by attempting to parse the value, and if it fails fallback to using the raw string. Unfortunately, while the happy path for time.Parse is fairly efficient, constructing errors is much slower and has a much bigger memory footprint.

We've developed fast iso8601 validation functions that cause no heap allocations to remediate this problem. We added a validation step to determine whether the value is a date representation or a simple string. This reduced CPU and memory usage by 5% in some programs that were doing time.Parse calls on very hot code paths.