All About that Base(64)

Base64 encoding is a strategy used in computing to transform raw data into readily-decodable, transmission-safe strings of text. Encoding results in bloated-but-perfect obfuscations of the original data.

For example, when we Base64 encode the string

“A human-readable string with a smiley :)”

it turns into:

“QSBodW1hbi1yZWFkYWJsZSBzdHJpbmcgd2l0aCBhIHNtaWxleSA6KQ==\n”

Why Bother?

Not all computer systems can correctly process all printable characters. Data loss occurs when one system sends information containing special characters – say, a smile emoji :) - to a system which cannot process those characters.

Base64 encoding attempts to prevent the loss or corruption of data when transmitting information between systems. It accomplishes this by translating source content into a format which ONLY uses a subset of 64 basic ASCII characters. Base16 and Base32 encoding strategies behave similarly, but employ smaller subsets of characters.

The base set of ASCII characters that are used for translation varies between different implementations.

Encoding Bloats Your Data

When encoded, a single character gets represented as multiple “base” characters – even if that character exists in the base set. As a result, Base64 encoded strings always grow to about 133% of their original data size (plus a little extra for the termination characters).

We can demonstrate this using Ruby’s Base64 module. Don’t forget to require the module in irb!

src_str = "The longer the source, the closer to an encoded size of 133% of the original!"
src_bytes = src_str.bytesize.to_f 
# => 77.0 bytes

enc_str = Base64.encode64(src_str)
enc_bytes = enc_str.strip.gsub('=','').bytesize.to_f 
# => 104.0 bytes

enc_bytes / src_bytes * 100 
# => 135 (percent)
 

In the above example, we’ve stripped the termination characters ==\n from the end of the encoded string before calculating its bytesize.

Base64 Encoding is NOT Encryption

Encoding a string results in something that looks like a jumbled mess of nonesense. But encoding a string is not the same as encrypting it!

Encoding only ensures data integrity. It does nothing for data security.

Decoding all that Insecure, Bloated Data

Encoding and decoding strings into Base64 is systematic; a patient, focused person could do both by hand if given a table of character translations. And enough time.

But the whole point of encoding data is to make it easily (quickly!) transmissable. Most programming languages have built-in functionality to deal with encoded strings.

Check out this blog post for examples of how to work with Base64 in Ruby.