Database Reference
In-Depth Information
|- corpus: ciphertext (required)
|- corpus_date: integer (required)
If you try to read the data outside of
ebq
, it will look like garbage. The
column names get rewritten to contain the encryption type. For example
corpus
becomes
p698000442118338_PSEUDONYM_corpus
. The values
themselves get rewritten as Base-64 encoded binary values;
hamlet
in
pseudonym encryption becomes something like:
ztYxwmeiiZB/
yDPC4W8u6g==
(depending, of course, on the encryption key).
Figure 13.3
shows what the
enc_shakes
table looks like in the BigQuery web UI.
Figure 13.3
Encrypted BigQuery table as seen in the BigQuery web UI
Encryption Modes
The encryption modes used in this example are pseudonym, homomorphic,
and probabilistic. These don't tell
ebq
which encryption algorithm to use—it
always uses standard AES. Instead of describing which encryption algorithm
to use, the encryption modes describe desired properties of the encrypted
data. The available encryption types are:
