SQL data warehouse and high-precision storage

Question

With HRD and BigTable, you have to deal with possible consistency for all queries that are not ancestral queries. Your code should be strong enough to cope with the results being outdated.

With the launch of Google Cloud Cloud on Google, they have stated a disclaimer: ( https://developers.google.com/cloud-sql/faq#hrapps )

"We recommend that you use Google Cloud SQL with High Replication App Engine applications. While you can use use Google Cloud SQL with applications that do not use high replication, doing so might impact performance." 

What does it mean? Does this mean that there are the same possible consistency issues using SQL with HRD? There is no concept of entity groups in SQL, but can this mean that specific SQL queries in certain circumstances provide obsolete results?

This would mean that the Oracle implementation of the SQL transaction contract would be broken and SQL would not function as relational database users would expect. If this is not the case, what is the problem with the fact that you have a master / slave or HRD model with SQL and why does Google give you the opportunity to choose a model with a lower performance?

Answer 1

(from the forum)

Cloud storage and data storage systems are independent. You can use one or both, as it suits your application.

We recommend using HRD applications because this type of application will be hosted with cloud-based SQL. Primary subordinate applications come from a different set of data centers where cloud sql has no presence. It will work, but it will be slower.

Answer 2

Quotes from the documentation:

"Google Cloud SQL is just an instance of MySQL that lives in the cloud. It has all the capabilities and functionality of MySQL."

To answer your question, the high replication / master -slave parameters for a relational database are not related to consistency, but to other factors, such as latency at peak loads and write access for scheduled maintenance. For a highly replicated data warehouse, latency is low, even if load spikes, and they are writable even when planning maintenance. Check out the comparison http://code.google.com/appengine/docs/java/datastore/hr/

And the second part of the question is why google will offer the master-slave option, which is not complete proof. The answer is that people who do not need full functionality and want to try GAE can use it.