The advent of the cloud and the introduction of Oracle Autonomous Database Cloud presents opportunities for every organization, but what's the future role for the DBA? This presentation explores how the role of the DBA will continue to evolve, and provides advice on key skills required to be a successful DBA in the world of the cloud.
Don’t believe us? Let’s look at the statistics
72% of IT Budget is spent on Generic Maintenance tasks vs Innovation
85% of security breaches occurred after the CVE was published
91% experience unplanned data center outages
And what’s scariest of all is that 80% of the outages experienced today are due to human error
In order to help change these statistics Oracle has introduce the Autonomous Database, a new category of cloud server which automates the complete lifecycle management of an Oracle Database with the help of Machine Learning.
So, how does the Autonomous Database change things for you?
5
Both ADW and ATP share the Autonomous Database platform of Oracle Database 18c on our Exadata Cloud infrastructure.The difference is how the services have been optimized within the database. When you start loading data into the autonomous
database, we store the data in the appropriate format for the workload.
If it is ADW, then we store data in columnar format as that’s the best format for analytics processing
If it is ATP, then we will store the data in a row format as that’s the best format for fast single row lookups
Query optimization: for analytics workload, we automatically parallelize the query execution to access large volumes of datain a short amount of time to answer biz questions If it is a transaction processing system, then we will automatically detect missing indexes and create them for you. Regardless of the workload we need to keep optimizer statistics current to ensure we get optimal execution plans. With ADWwe are able to achieve this by gather statistics as part of all bulk load activities. With ATP, where data is add using more traditional insert statements statistics are automatically gathered periodically.As the data volumes change, or new access structures is created, there is the potential for an execution plan to change and any change could result in a performance regression so we use Oracle SQL Plan Management to ensure that plans only
change for the better.
Automatic Indexing is an expert system that implements indexes based on what a performance engineer skilled in index tuning would do. But unlike a human it is able to work 24 hours a day, 7 days a week, 52 weeks a year. It also takes full responsibility for its decisions and hence its decisions are validated before they are implemented to ensure proposed indexes really do improve performance. The entire process is fully transparent to the DBA via a detailed report of what happens each time automatic indexing kicks in.
It’s a simple 6 step processing:
1.CapturePeriodically captures SQL statements into a SQL repositoryInclude plans, bind values, execution statistics, etc. (AWR for SQL only)
2.Identify CandidatesIdentify candidate indexes that may benefit the newly captured SQL statementsCreates index candidates as unusable, invisible indexes (just metadata)
3.VerifyAsk the optimizer if index candidates will be used for these statementsIndex candidates not used by the optimizer are automatically droppedComplete creation of chosen indexes and run captured statements to verify that the indexes did improve performance
4.ConfirmIf performance is better for all statements, the indexes are marked visibleIf performance is worse for all statements, the indexes are droppedIf performance is worse for some, the indexes are marked visible except for statements that regressed
5.Validation Each UseThe validation of the new indexes continues for other statements, onlineFirst session that runs each affected SQL validates benefit, and avoids index if none
6.MonitorIndex usage is continuously monitoredAutomatically created indexes that have not been used in a long time will be droppedRebuilds decaying indexes
10
Highlight Findings related to:
Oracle Best Practices
CIS Oracle Database Benchmark
EU GDPR
What does DBSAT Check?
Security Configuration
Data Encryption
Auditing Policies
Fine-grained Access Control
Database and Listener Configuration
OS File permissions
Security Patches
Users and Entitlements
User Accounts, Privileges and Roles
Sensitive Data
Which type, where, how many - To discover sensitive data in the database, DBSAT looks for column names and column comments.
The business is going to look to you to determine what database cloud service they should use
You will be in charge of, and in control of, the end-to-end service levels
You need to know what each Database Cloud Service offers you in terms of
Availability
Security
Performance
Scalability
Take advantage of trial accounts to get familiar with provisioning and using a cloud environment
Oracle connection pools, for example, use FAN to receive very fast notification of failures, to balance connections following failures, and to balance connections again after the failed components are repaired. So, when a service at an instance starts, the connection pool uses the FAN event to route work to that resource, immediately. When a service at an instance or node fails, the connection pool uses the FAN event to immediately interrupt applications to recover. FAN is essential to prevent applications from hanging on TCP/IP timeouts.
Application Continuity masks outages from end users and applications by recovering the in-flight work for impacted database sessions following outages. Application Continuity performs this recovery beneath the application so that the outage appears to the application as a slightly delayed execution.
Back to Penny
In step 1, we create a ML Model, ‘season_tks_model’, which classifies the likelihood of fans purchasing a season ticket based on detailed history – which would typically reside in ADW. (the settings table is there to decide what to to when certain columns are null – for example).
In step 2, we apply the model to determine the probability of fans most likely to renew their season ticket. In the example shown, we want to know the probability of a 40-year old married person with $3,500 in the bank will purchase. By applying ML classification model (in this example), we can more readily identify top prospects for a sale from a long list of names/leads.
Has standard GoldenGate restrictions or rowids, nested tables, identity columns, etc.
ADW Schema Advisor provides guidance on Oracle Schemas to be migrated to ADW
Analyzes only the Schema objects, not the workload
Generates an easy to understand report
Easy to use, fast, command-line tool
Analyzes the metadata, does not depend on Data size