Base64 Encoding Overhead
While Base64 is a popular way to encode raw byte data and send it over a connection that can only handle text, i wonder about its efficiency. Just how much overhead is involved when I convert by bytes to a Base64 encoded string?
In this article I run some tests to see what kind of overhead I can expect when using the Base64 encoder provided in the .NET framework
References
- The Base16, Base32, and Base64 Data Encodings (rfc4648) [tools.ietf.org]
- Base64: What is the worst possible increase in space usage? [stackoverflow.com]
- What is the purpose of base 64 encoding and why it used in HTTP Basic Authentication? [stackoverflow.com]
- How do you embed binary data in XML? [stackoverflow.com]
How I tested
To test the practical effects of Base64 encoding I created C# strings of 1, 10, 100, 1000 and 10000 characters in length. I converted each string into a byte array (using System.Text.Encoding.ASCII.GetBytes(SringHere) ) and then encoded the byte array into base64 (using Convert.ToBase64String(ByteArrayHere) ).
Results
| String Length | Byte Array Length | Base64 String Length | Raw Increase (in Bytes) | Overhead |
| 1 | 1 | 4 | 3 | 300% |
| 10 | 10 | 16 | 6 | 60% |
| 100 | 100 | 136 | 36 | 36% |
| 1000 | 1000 | 1336 | 336 | 33.6% |
| 10000 | 10000 | 13336 | 3336 | 33.36% |
From the data I've collected it appears that the overhead associated with using Base64 approaches 33 1/3% the more data you choose to encode.