Choosing a database
Two questions choose among five databases
Relational: Cloud SQL, AlloyDB or Spanner. Non-relational: Firestore or Bigtable.
Google Cloud has five databases in this exam, and two questions separate them. First: is the data relational, in tables, queried with S Q L? If yes, and you want a familiar engine in one region, Cloud S Q L. If it is PostgreSQL with heavy demands, AlloyDB. If it must span the world and stay consistent, Spanner. If not relational: documents for web and mobile apps go to Firestore. And huge volumes with fast, simple lookups go to Bigtable.
Choosing a database · Going deeper · 1 of 2
The seven data products
Six store data for applications. BigQuery is the one built for analysis.
Choosing a database · Going deeper · 2 of 2
Relational, non-relational, object
SQL for relational tables, NoSQL for the rest, object storage for whole files.
Cloud Storage classes
Colder storage is cheaper to keep and dearer to read
From Standard to Archive, keeping 1 TB falls from $20 to $1.20, and reading it rises from $0 to $50.
Cloud Storage keeps files of any kind, in four classes. Watch both columns: the cost to keep a terabyte for a month, and the cost to read it back. Standard: the most to keep, and free to read. For data in constant use. Nearline: half the price to keep, and a fee to read. For about once a month. Coldline: cheaper again to keep, dearer to read. About once a quarter. Archive: almost nothing to keep, the most to read. Less than once a year. And if you cannot predict the pattern, Autoclass moves each object for you.
Cloud Storage classes · Going deeper
The classes and their shortest stay
Nearline 30 days, Coldline 90, Archive 365. Autoclass decides for you.
Check yourself
Each mark is a time the backups were read during one year. Which storage class is designed for this pattern?
Analytics and BI
BigQuery finds the answer, and Looker shares it
BigQuery is the serverless warehouse. Looker defines each measure once and delivers it to everyone.
Storing data is half the story. Here is how it reaches the people who decide. Data arrives from many places: sales systems, the web site, even spreadsheets. It lands in BigQuery, one warehouse that is asked questions in S Q L, with no servers to size. Looker sits on top. It gives each measure, such as revenue, one agreed meaning. And it delivers the results as live dashboards, scheduled reports, and analytics inside other applications.
Analytics and BI · Going deeper
What Looker adds
Self-service access, one definition of each measure, and real-time reports.
Streaming analytics
A stream is analysed while the event still matters
Pub/Sub carries events and Dataflow processes them: 5 seconds, against 11 hours for a nightly batch.
Some data loses its value within minutes. That is the case for streaming. Events arrive all the time: payments, clicks, sensor readings. Pub/Sub receives them and delivers each one to whoever subscribes. Dataflow processes them as they flow: cleaning, joining and totalling. BigQuery stores the result and answers questions. And something acts on it, such as an alert. Take a stolen card used at two minutes past two. Streaming knows in seconds. A nightly batch knows almost twelve hours later.
Streaming analytics · Going deeper
The pipeline products
Pub/Sub carries, Dataflow transforms, Managed Service for Apache Spark runs existing Spark work.
Check yourself
Which product belongs at the "?" in this pipeline?
Moving databases
Keep the engine and move, or change the engine and gain
Migrate to the same engine on Cloud SQL, or modernise to AlloyDB or Spanner.
The last objective is about databases you already have. To migrate is to keep the engine. A My S Q L database in your datacentre becomes Cloud S Q L for My S Q L. Little changes, it is done in weeks, and Google takes over patching and backups. To modernise is to change the engine: an ageing commercial database becomes AlloyDB or Spanner. The application must change too, but you gain scale and shed licence fees. In both cases Database Migration Service copies the data while the old database keeps running.
Moving databases · Going deeper
Clue words for data products
Match the phrase to the product: twelve pairs cover section 2.
Check yourself
A game studio needs the store marked "?". Which product fits?
Lesson complete
You finished “Data products and analytics on Google Cloud”
Answer the questions along the way to see your score here.