Google's cloud spanner

1
Google's Cloud Spanner
2
The beat launch of Cloud Spanner was declared
today by Google and it is a new globally
distributed database service for the purpose of
mission applications. There are other cloud-based
database services that are joined by Cloud
Spannerlike Cloud SQL, Bigtable and the Cloud
Datastore but with the main difference of
providing developers the best relational
databases and NoSQL databases and it is
transactional with easier scalability. Cloud
Spanner is another solution for developers who
are striking the limits of cultural relational
databases like PostgreSQL or MySQL. Google
usage of this version internally is the reason
behind the name Cloud Spanner and a paper was
published in the year 2012. In the year 2007,
Google started working on Spanner just to search
for an alternative to MySQL which ran lots of
company products.
More...
3

• Spanner is not a complaint of MySQL or ANSI SQL
  and there is no such plan
• Moving workloads from MySQL is the target for
  Google's Spanner and is the biggest interest and
  use case. Most of them felt bad about the lack of
  MySQL support both in the present and in the
  Spanner roadmap. There is a need for more
  pressure from MySQL to migrate or move a workload
  and for ensuring a performant implementation.
• Full RDBMS functionalities are not offered by
  Cloud Spanner similar to referential integrity.
• A parent or child relationship is needed by the
  Spanner when designing your data model to define
  at the table level. It is said that in other
  places this is by choice. There is no referential
  integrity performed on an automatic basis if you
  do not define tables as interleaved. If you later
  believe that you change the table into a child
  then you are going to have a drop or recreate the
  table.

More...
4

• The online schema does not offer cloud spanner
  for changing the keys.
• Drop and replace of the table will be present in
  you and this happens when you need to change the
  PK of your table. This implies that you need to
  pause your workload and it means to take a
  downtime. And you need to understand why you want
  to change the table's PL? In bold or all caps and
  red the spanner documentation is highlighted.
• Index choosing is difficult when cloud spanning
  comes into the picture and it recommends using
  FORCE_INDEX directives in your code.
• SQL operation is done from SQL to Google's APIs
  and the efforts of the application programmer
  will have to be supported by data architects and
  they are people who are aware of both design
  around the limits of Spanner along with
  neglecting Spanner-specific anti-patterns which
  reduce performance. 

More...
5

• This implies changing Pks to hashes and taking
  away auto-increment and correctly assigning
  parent/child data relationships as interleaved
  tables and interleaving needed indexes along with
  design time.
• Join DBA Course to learn more about other
  technologies and tools.
• Stay connected to CRB Tech for more technical
  optimization and other updates and information.