For this chapter, although there are no technical requirements, if you are familiar with the Oracle Cloud Infrastructure (OCI) Console, you will be able to visualize the topics discussed.

Before we start learning about ADB, let’s first understand what OCI is.

OCI was designed to satisfy the needs of enterprise workloads that often require high performance, security, elasticity, availability, and integrity for their critical applications. Enterprises today want to lower their cost and move from a traditional CAPEX-based model to an OPEX-based culture. At the same time, they need a rich set of cloud services and automation capabilities built using cloud-native technologies to provide a comprehensive cloud solution for customers. OCI provides services around infrastructure, data management, analytics, applications, development/DevOps, governance, and security to cater to requirements from big to small enterprises. OCI is not just limited to Oracle’s data center but can also be extended to customers’ data centers; an offering called Cloud@Customer, which runs behind the company’s firewall, is available, and solves data sovereignty requirements. OCI is also available as dedicated regions for those workloads that require an in-country location or have data sovereignty requirements.

OCI’s ADB is a self-driving, self-securing, and self-repairing fully-managed database environment available on the cloud as well as on-premises. As of right now (the cloud is all about change), four distinct workload types are available with autonomous databases: Autonomous Transaction Processing (ATP), Autonomous Data Warehouse (ADW), Application Express (APEX), and Autonomous JSON Database (AJD).

You can build data-driven apps and gain operational insight in real time without worrying about the operational aspects of a database in terms of maintenance tasks such as backups, patching, upgrades, and performance tuning. You can scale the number of CPU cores or the storage capacity of the database at any time without impacting the availability or performance of the database system. With cloud-native developments and auto, we have highlighted and discussed auto scaling in detail in other chapters. Here, we give an overview of all the scaling features OCI provides and its automated responses to your workload needs.

ADB is built upon a very solid foundation with more than three decades of technical innovations developed by Oracle, providing customers flexibility combined with Machine Learning (ML) and Artificial Intelligence (AI). Oracle manages everything for you, so you can focus on your data, development, and delivering solutions that impact your business.

ML models and algorithms run inside Oracle ADB. It brings the following advantages:

• Data stays in place
• Massive parallel execution
• Flexible model building

In addition to these benefits, ADB also supports key Oracle database features and open source programming languages:

• Structured Query Language (SQL), R, or Python
• Oracle Data Miner (ODM)
• Oracle AutoML

With Oracle Machine Learning and Oracle ADB, users have a variety of options for building and deploying models involved in data science projects, whether they use in-database algorithms or open source Python algorithms. An autonomous database uses AI and ML to achieve a complete, automated provisioning experience, applying security, automated patch updates, continuous availability, and performance tuning based on the workload types; managing changes; and avoiding mistakes. Oracle Machine Learning for Python (OML4Py) in Oracle ADB upholds versatile in-dataset information investigation and arrangement utilizing local Python grammar, conjuring in-dataset calculations for model structure and scoring, and implanting the execution of client-characterized Python capacities from Python or REST APIs. Likewise, OML4Py incorporates the AutoML interface for automated calculations and component choice and hyperparameter tuning to augment model execution. On the other hand, ODM, which is an extension of Oracle SQL Developer, helps develop ML methods.

Let’s discuss a bit about security in autonomous databases. We will go through the details of it in Chapter 7, Security Features in Autonomous Database. All data in ADB is encrypted, and users or applications need to be authenticated in order to use the database. ADB does not require any manual configuration for providing encryption – whether data is at rest or in motion, all connections use certificate-based authentication over Secure Socket Layer (SSL). ADB enforces strong password complexity for all users based on Oracle Cloud Security standards. ADB provides a network Access Control List (ACL), using which databases can only accept connections from allowed IP addresses and reject all other client connections. ADB also provides network access through private endpoints that help organizations implement strict security mandates to only allow connections privately from inside a Virtual Cloud Network (VCN), and traffic never uses public subnet and public internet within your tenancy VCN.

Quick note

In OCI, ADW was the first offering launched in 2018. Later, ATP was added to the service portfolio of offerings at beginning of 2019. Recently, in August 2020, Oracle also added JavaScript Object Notation (JSON) databases to the Autonomous Database service catalog, known as AJD.

Quick note

You can think of a VCN as a private network set up inside an Oracle data center, which consists of several firewall rules and communication gateways. The components of a VCN are one or more subnets, Internet Gateways (IGWs), Dynamic Routing Gateways (DRGs), route tables, security lists, and DHCP options. When you create a VCN inside OCI, most of these components are created by default. Another thing to keep in mind is that a VCN covers a single, continuous IPv4 CIDR block of your choice. In other words, you can say that a VCN provides software-defined networking in OCI.

Always Free ADB

Always Free ADB is available through Oracle’s Free Tier, which provides customers with up to two instances of ADB (Serverless/Shared) for every tenancy. Always Free ADB supports both ATP and ADW workload types. Customers can upgrade a Free database to Paid anytime.

Quick note

You can sign up for an Oracle Free Tier account by navigating to https://www.oracle.com/cloud and clicking on the Try Oracle Cloud Free Tier button on the right side.

The key characteristics are as follows:

• It has a fixed configuration: 1 OCPU, 20 GB of storage, and 8 GB of memory
• Up to two Always Free instances in every tenancy’s home region
• Most ADB functionality available, except Scale Up/Down, Storage auto scaling, Update License Type, Manual backup, and Restore
• Upgrade an Always Free database to Paid anytime

Always Free ADB gets automatically stopped after 7 days of continuous inactivity. After 90 (cumulative) days of continuous inactivity, Free ADB instances are also automatically terminated. Users are notified via console UI banners for both of these events.

Always Free autonomous databases can only be created in your account’s home region as shown:

Figure 1.1 – The Always Free ADB option during deployment

Customers can create autonomous databases quickly and easily using the OCI Console, Command-Line Interface (CLI), Software Development Kit (SDK), and Terraform. To create a database, customers can log in to their OCI Console and select Oracle Database | Autonomous Data Warehouse | Transaction Processing | JSON Database. You need to provide details such as the database name, the number of OCPUs, storage (in TB), and the admin password. In a couple of minutes, a fully ready autonomous database is ready for use by the customer. Customers can perform various management operations on their databases, such as starting, stopping, restarting, backing up, cloning, using Data Guard, and monitoring. Backups are automatic and the customer has the option to take a full backup anytime, as well as the ability to restore to a “point in time” backup. Backups are retained for 60 days by default and the customer can configure it to be more or less. The customer can scale their database CPUs and storage without any downtime. Using administration credentials, customers can access and start using the ADB service using a separate service console. You can also update admin credentials anytime.

Technically, ADB is built on OCI Exadata infrastructure. Each ADB database is an independent Pluggable Database (PDB) to which the customer doesn’t have host access. Oracle manages the entire life cycle activities of the database based on customer inputs and preferences. You can check the ADB page within Oracle Cloud Console as depicted in the following screenshot. It shows deployed ADBs within a region.

Figure 1.2 – ADB page on the Console

As we can see in Figure 1.2, a single page has both a shared and dedicated infrastructure link for the easy creation of these services and navigation capabilities.

Let’s see what makes an Oracle database autonomous. We will look at various building blocks for autonomous databases. You will notice that starting from Oracle Database version 9i, Oracle introduced several automation capabilities around memory management, workload monitoring, and self-tuning capabilities, which set the base for autonomous databases. With the acquisition of Sun Microsystems, Oracle drove a database infrastructure with engineered systems focused on more automation capabilities and bringing data processing to the storage layer, with innovations such as Smart Scan, query offloading, a storage index, columnar compression, and so on. These database platforms are preconfigured and highly optimized for running database workloads, pre-tested across thousands of deployments, thus forming the base for autonomous databases.

The ADB building blocks are as follows:

• Oracle Database Enterprise Edition (DBEE)
• Oracle Exadata Database Machine
• OCI
• ML
• Oracle’s best practices
• Oracle’s knowledge base

We will talk about each block in detail in the next sections.

Oracle DBEE

If you have prior knowledge of Oracle databases, you will already know that Oracle had two distinct editions of databases targeted for different market segmentation: a Standard edition and an Enterprise edition. As the Enterprise edition was built to suit the high-performance requirements of enterprise customers for transactional and analytical workloads, it has several features that make it enterprise-class. With traditional database deployments, the DBA needs to tweak several configuration parameters based on workload types, not just the database but also the Operating System (OS) and network configuration – everything that goes with any production-ready database deployment. ADB removes these complexities and comes preconfigured with optimal values based on deployment types.

Oracle DBEE sets the foundation for autonomous databases. The database options available with DBEE provide the required capabilities to run ADB. The following options give ADB autonomous capabilities:

• Real Application Clusters (RAC): Provides high availability functionality, including scale-out architecture, failover in case of instance failure, and online patching to avoid downtime
• Active Data Guard: Provides standby capabilities and is used for disaster recovery purposes
• Parallel SQL: A core feature for prioritizing SQL’s parallel degree based on system resources and policies
• Multitenant option: Provides the required functionality for Agile development
• Database In-Memory: Provides high performance for analytic queries
• Transparent Database Encryption (TDE): Part of the Oracle Advanced Security options – a default for data encryption
• Database Vault: Segregation of duties running within a database kernel – used for compliance requirements and blacklisting and whitelisting of users and programs

Quick note – database parameters

All database parameters are set to optimal values based on workload type. Users can only change a limited number of parameters. These parameters are shown in the following screenshot.

Figure 1.3 – List of allowed parameters for modification in ADB

As we can see, most of the changeable parameters are around the user’s profile, related to NLS, time zones, and so on. Oracle sets all other parameters to optimal values by default.

Oracle Exadata Database Machine

If you are new to Exadata Database Machine, you can consider it a combination of software and hardware optimized to run Oracle Database. The current version of Exadata is X9-M and it will be keep being updated based on the latest version. Normally, Oracle follows a cycle of 12 to 18 months to release a new generation of Exadata machines. The very first version of Exadata was released back in 2008. With each Exadata refresh cycle, customers get the most recent CPU processors, memory, increased disk capacity, flash, and high-speed networking components, which provide increased performance, security, availability, and management capabilities. Exadata is known as a great consolidation platform because of the massive capacity and performance available with these machines.

Oracle introduced a storage layer within the database machine, with several innovations supporting scale-out architecture, and parallel query operation, which greatly optimized data processing at its storage layer. Exadata solved two major problems: avoiding network bottlenecks for data movement within the machine through SQL offloading and, at the same time, providing a larger network bandwidth (100 Gbit/s) Ethernet fabric for data access. Exadata also provided separate Ethernet ports for data center connectivity and management operations such as backups. Some of the key innovations within machines can be considered Smart Scan, query offloading to storage, storage indexes, flash caching, resource management, Hybrid Columnar Compression (HCC), and in-memory database capabilities with fault tolerance.

Quick note – ADB platform

ADB runs on RAC on Exadata. ADB decides where to place each database during provisioning. A fewer number of instances are preferred when possible. Even though it’s running on RAC, the database can only be open on one node.

OCI

OCI provides required technologies such as networking elements, VCNs, subnets, virtual firewalls (network security groups), security lists, communication gateways, identity and access control, automated provisioning, logging, audit, monitoring capabilities, and so on, which are needed to run Exadata Cloud Service natively. OCI provides end-to-end security with a focus on the unified, automated, prescriptive security experience that makes life easier for customers. Identity management is a key focus for OCI, which helps simplify a customer’s security landscape, starting with data and then moving through the infrastructure, network, monitoring, and edge services. At the data or database layer, OCI supports encryption at rest and in transit and supports hardware security modules.

Within infrastructure, for compute instances, OCI supports hardened OS images, autonomous Linux, hardware root-of-trust, and signed firmware. In the networking domain, OCI supports isolated network virtualization with off-box Network Interface Cards (NICs), private networking with FastConnect, and security zones, which can be used to apply context-specific security policies to compartments. For monitoring, OCI has integrated Cloud Security Posture Management with Cloud Guard. Cloud Guard is very dynamic and OCI releases new services every week for customer needs. Recently, Oracle announced Scanning Service, which scans compute hosts and container images for vulnerabilities. The Bastion service automates the configuration of secure Bastion servers. The Certificates service automates the provisioning and management of private and public certificates. Threat Intelligence Service centralizes threat intelligence and vulnerability feeds integrated across cloud services.

Oracle best practices

Every organization emphasizes adopting best practices, Oracle has published several “best practices,” which are based on expert recommendations for deploying a product, fine-tuning, configuration changes, and so on. In addition to this, Oracle’s Maximum Availability Architecture (MAA) focuses on best practices for the availability of applications based on the categorization of Service-Level Agreements (SLAs). Oracle has a set of best practices around security called defense-in-depth. Best practices allow the Oracle database to run with optimal efficiency. Oracle uses several of its core features to provide the required level of optimization; using technologies such as online reorganization allows online operation for table redefinition without compromising the availability of the system. Using Resource Manager, Flashback Technology, Application Continuity, and Transparent Application Continuity protects against several kinds of failures. RAC protects against node failures, and this also enables rolling patches, service draining, and zero-downtime planned maintenance. Application Continuity protects transactions from failures, allowing safe transaction replay using a JDBC replay driver and Transaction Guard. Using best practices ensures that Oracle technologies achieve the highest level of performance, availability, and security.

Oracle knowledge base

Oracle has a knowledge base of several years built from a diverse set of customer issues, whether service requests, product management contributions, development experiences, bugs reported by customers, or enhancement requests for features and services. The Oracle system for support tickets, known as MOS, is an interface for Oracle and customers that allows them to open up support tickets in case they need help. Customers can visit MOS to explore several knowledge base articles, how-tos, and so on. As part of problem diagnostics and resolution, customers often provide logs, screen shares, diagnostic collection, OS details, trace files, and so on. These files are a great source of information for Oracle for using these as input to build AI and ML for intelligent data analysis and problem-solving.

ML

ML is an important function in the autonomous database, where the database uses ML algorithms and automates the most important aspects of the database, such as security, database backups, patching, performance tuning, index creations, and several routine tasks that are typically performed by a DBA. It has increased productivity, as no human intervention is needed, thus freeing up time for other inventions. OCI also provides an ML platform as a service that can be used by customers to implement their own solutions while using ADB as the database of choice. ML is also a set of tools available in Oracle Cloud that customers can use to implement their own solutions.

We have learned about the different building blocks for ADB, so now it is time to look at the classification of ADB based on workload characteristics. OCI has tailored these databases based on types such as DWH, OLTP, and JSON.