What's the general schema that Gmail uses with Bigtable

Question

Google now allows you to develop apps using BigTable (hosted as a product called "Cloud Bigtable" in Google Cloud Platform). However, I can't find too many example on how to design a schema for it. They have a document, but it's very high level: https://cloud.google.com/bigtable/docs/schema-design

My Question: what's the approximate schema for gmail (just the email list component). Is it a tall or wide design? What do they use for primary keys?

Any other production examples from big apps would be appreciated, but I think Gmail would be a great example.

Answer 1

Schema design is very specific to your application. HBase fortunately has quite a bit of public material to help:

• There is a Quora Thread on production schemas.
• The Time Series material that comes after the doc you mentioned is good.
• The HBase documetnation has a chapter on Schema design.
• Lars George's HBase book has some good discussion. He also talks on YouTube
• Ian Varley's HBaseCon talk covers this as well.
• Amandeep Khurana has an article.