Nodejs and browser based JavaScript differ because Node has a way to handle binary data even before the ES6 draft came up with ArrayBuffer. In Node, Buffer class is the primary data structure used with most I/O operations. It's a raw binary data that is allocated outside the V8 heap and once allocated, cannot be resized.
Before Nodejs v6.0, to create a new buffer you could just call the constructor function with new keyword:
https://gist.github.com/ea236e5343fc0e4a5295071d27fa48d3
To create a new buffer instance, in latest and current stable releases of Node:
https://gist.github.com/909e075f2066c5826d0e60de85fbe831
The new Buffer() constructor have been deprecated and replaced by separate Buffer.from(), Buffer.alloc(), and Buffer.allocUnsafe() methods.
More information can be read through official documentation.
Buffers can convert to JSON.
https://gist.github.com/15a3d9b83a95ef426b05ddca19dbb374
The JSON specifies that the type of object being transformed is a Buffer, and its data.
https://gist.github.com/1c42f6674af95b8d347338b7546669d1
https://gist.github.com/ed7997521d3147c252194c0ba4074a0b
.toString() is not the only way to convert a buffer to a string. Also, it by defaults converts to a utf-8 format string.
The other way to convert a buffer to a string is using StringDecoder core module from Nodejs API.