Oracle Cloud Infrastructure

Oracle is consistently on the top 10 lists of cloud service providers because of its Infrastructure-as-a Service (IaaS) provider approach. With more than 40 years in tech, Oracle has learned how to leverage its extensive knowledge of evolving database technologies, wealth of applications, and trial-and-error of early infrastructures.

They have established a goal of reaching 20 different regions to provide faster, more secure cloud access in more locations. All of their regions follow the same structure but provide the necessary controls for businesses to customize their cloud storage space as needed. In addition to expanding availability around the world, Oracle has worked to address the four primary perceived barriers that prevent businesses from making the transition to the cloud:

1. The cloud is complex, and companies don’t know how to map the move.

2. Companies lack the dedicated, knowledgeable IT staff to make the migration.

3. It isn’t adequately secure.

4. Working in the cloud is slower and will reduce performance.

In response to these concerns, Oracle released the second generation Oracle Cloud Infrastructure at the end of 2019. The new generation addresses these concerns by establishing Oracle’s four pillars: security, enterprise expertise, cost, open implementation. Many of the initial issues with migration and implementation were either significantly reduced or eliminated. 

This article provides an overview of the structure, a look at what makes Oracle’s cloud infrastructure different from other major IaaS providers, and details the stated focus Oracle places on its four pillars.

Understanding Oracle Cloud Infrastructure

The focus of Oracle’s second generation cloud infrastructure is to enable enterprise services with the option of tight x and y controls. Businesses need cloud solutions that more closely mirror their current internal structure. The second generation provides a robust structure with an eye for future scalability, including the following:

• Agility

• Control

• Innovation

• Performance

• Security

• Speed

Oracle combines the utility and elasticity offered by the cloud with the control and performance available with on-premises infrastructure. For example, the Identity and Access Management service controls the type of access groups of users have to which network resources. Oracle uses a simple SQL-style syntax to make it easier to create and manage compartments and policies within the IAM service.

Oracle controls are easy to use because companies can create a cloud hybrid with their existing system. While the servers remain on-site, Oracle cloud hybrid technology allows businesses to create automatic backups to the cloud that are fully scalable, as well as backups on demand. On-site backups to tape can still be executed on a schedule in addition to cloud-based data backup and retrieval.

Infrastructure Overview

Customers select the services they want to implement. Oracle provides them the secure platform with high-performance capabilities, including the following:

• Viewing and monitoring database cloud services

• Hybrid cloud management workflow

• The ability to define Oracle Cloud database service settings 

• Scalable storage capacity

This quick overview provides an overview of how Oracle established a world-wide cloud without getting too technical. 

Worldwide Availability and Availability Domain Structure 

As of November 2019, Oracle has 16 regions providing world-wide services, with plans to build another 20 regions by the end of 2020. Each region is composed of seven layers:

• Regions and Realms

• Availability Domains

• Fault Domains

• Physical Layer

• Virtual Layer

• Services

• Governance

Regions and Realms 

A region has 1 or more availability domains, typically designed with 3 availability domains. Each of these has 3 fault domains. Regions are all independent of other regions and can be separated by vast distances. Sharing data across multiple availability domains and across regions reduces the risk of outages due to large weather systems or events. Regions are further grouped into realms, and anyone using a single region in a realm has access to the resources of all regions within that realm. This gives individual users faster remote access to resources.

Availability Domains

All Oracle regions are connected through a low-latency dedicated network backbone. Each infrastructure region has at least 3 availability domains composed of physical data centers located within 30 to 50 miles of each other. This architecture provides geographic isolation, significantly reducing the risk of physical simultaneous failure while improving security. Each availability domain typically has at least 3 fault domains. Each availability domain within a region provides high-availability connectivity and the ability to replicate systems in multiple availability domains for disaster recovery and high availability. This is possible because availability domains are connected by a low-latency, high-bandwidth network.

Fault Domains

Fault domains work within a single availability domain to give granular control of application availability. Networked computers are distributed across physical hardware within the availability domain, providing greater fault tolerance. As these individual resources are pulling from several different hardware devices, it creates redundant power supplies to prevent power faults due to power outages. This keeps customer data, databases, and apps protected and available while giving customers the ability to perform repairs or maintenance without interruptions.

Physical Layer

The Oracle cloud is built on a physical layer of hardware networks and virtual networks. Because Oracle implemented a flat network design, customers will have the fewest number of hops to reach a different server. It also covers all servers, maximizing customer bandwidth (25Gbps). The physical layer is made up of all the data sources used to process multiple data source queries.

Virtual Layer

Oracle virtualizes input/output (IO) in the network (instead of through a virtualized software layer), providing better isolation and encapsulation. This is called “off-box” virtualization because it takes network and IO virtualization off the server box and puts it in the network. Off-box virtualization decreases the load on the server and speeds up access to data.

Customers can create virtual machines, or they can access to bare-metal compute instances. All of this is on the network so that customers can access everything through a single portal and API set. One of the primary benefits to the bare-metal compute instances is that they give customers the same level of control that customers have on their current server. 

There are different virtualization interfaces that can be used to access virtualization services including Oracle VM VirtualBox, Microsoft Hyper-V, and VMware Vcenter. Each of these can be used to manage Oracle VDI.

Services

Many businesses initially use the cloud to store a vast amount of data and apps. Oracle offers several cloud services to meet a variety of needs including the following:

The full list of available Oracle cloud infrastructure services is here. 

Governance

Compartments are in the different regions to keep cloud resources organized across the availability domains. Each cloud resource is located in a single compartment, and the company administrator grants permission to access different compartments. Fine-tuning access is essential for remaining compliant with rules and regulations and to establish internal standards. The first generation of the cloud largely lacked this level of control. 

Best Workloads to Run on Oracle Cloud Infrastructure

While Oracle provides a platform to easily run and maintain any workload, Oracle Cloud Infrastructure performs particularly well for these purpose-built workloads: 

Enterprise Database Applications

Whether the customer’s enterprise database application is one of Oracle’s own (such as Oracle Database or Fusion Middleware), from a third party (such as JD Edwards World or Autovue), or a custom app, the Oracle cloud is designed to handle the many demands of the app. This is because enterprise database apps can leverage the backend Oracle Database.

High-Performance Computing

Dedicated High-Performance Computing can be a major budgetary sink to an organization. By utilizing the Oracle cloud, infrastructure can be provisioned to meet the requirements for simulations, digital twins (a digital replica of a physical object including physical assets or systems), and risk modeling, and customers only pay for the computing power they use.

A new high-performance computing instance can be established within minutes, compared to the hours that it takes to create a hardware instance. With the 25 Gbps network of the bare-metal instance, the data for these workloads moves as quickly as the data for smaller, simpler projects. Oracle accomplishes this feat by using a single-tenant cloud (instead of the usual multi-tenant), which reduces hypervisor overhead and performance variability. This streamlines the movement of data. Oracle can also provide higher bandwidth and lower latency by offering higher clock speeds and RDMA-based cluster networking.

Big Data

Despite the benefits, many businesses have avoided big data to some degree because it isn’t possible to house it on local servers due to the vast volume of data. The Oracle cloud infrastructure allows for the quick provisioning of a new framework and the ability to scale it as needed, giving customers the benefits of big data from a reliable, secure location. Oracle’s second-generation cloud infrastructure is different from other solutions as it offers a range of fully automated options.Oracle cloud infrastructure can self-patch and self-manage, which eliminates some forms of human error and improves security. 

Oracle is also unique in its layered approach to the cloud. They use an application layer (for housing a company’s applications), which sits on a platform layer that is powered by the Oracle Autonomous Database. The Oracle Autonomous Database is a family of self-securing, self-driving, and self-repairing cloud services. The platform layer is also where the integration services reside that connect it to the application layer and the infrastructure layer. This allows Oracle second-generation cloud services to handle more workloads at faster speeds.

Cloud Native Application Development and Hosting

Cloud-native applications are applications that have been developed using cloud-based technologies. Cloud-native apps are managed and hosted in the cloud. These kinds of apps offer scalability that customers control, but they can be time-consuming to create and maintain. The primary problem with them is that traffic varies. Any event that drives an unexpectedly high volume of traffic to the app can crash it, upsetting both staff and customers. 

Oracle’s cloud minimizes the risk by providing the necessary support to keep these apps running. Acting as a host for the apps, apps are set up on bare metal, a virtual machine, or a container. Once it is set up, they can be scaled. Oracle utilizes several tools to achieve these remarkable milestones: containers, microservices, and streaming.

Containers

The Container Engine for Kubernetes is a service that uses Kubernetes clusters with Cloud infrastructure to achieve faster deployment, portability, scalability, and environmental stability.

Microservices

The use of microservices enables different teams within a company to build and test different parts of an app being developed simultaneously. 

Streaming

Streaming allows a large volume of data to flow from individual devices and web applications and be automatically provisioned while minimizing deployment and management.

Other Cloud-native services include continuous integration that is very useful for developers. The use of infrastructure as code reduces the amount of time required to deploy servers, storage, and networking. Serverless functions automates repetitive tasks.

Oracle’s Four Pillars of the Second-Generation Cloud Infrastructure

Launched in October 2019, Oracle’s second generation took the lessons learned from the first generation. Restarting from the beginning, they addressed many of the issues raised by customers of the first generation (also known as OCI classic) and concerns expressed by enterprise businesses about their journeys to the cloud. 

Oracle adopted four principles to guide the development, with each of these principles, or pillars, designed to address business requirements. 

1. Robust security measures
2. A focus on enterprise expertise
3. Minimizing cloud costs
4. An dedication to openness

Robust Security Measures

In 2019 alone, there were 3,800 publicly disclosed data breaches, resulting in the exposure of nearly 4.1 billion sensitive records. Most businesses consider data security the company priority. 

Oracle ensures that every phase of the transition to the cloud is as secure as possible. Oracle’s security-first approach to cloud architecture applies core-to-edge security. Oracle employs an “edge strategy” that resolves web asset names, ensures geolocation accuracy, and utilizes dynamic traffic steering to deploy a layer of security and performance for traffic coming from web applications. They use various forms of DDoS protection and a web application firewall, which provides multiple layers of security at the “edge” where users and devices connect to the network. 

These additional layers of edge protection include bot management, web app security, API endpoint protection, and web app access control to defend against targeted attacks. To protect the cloud from volumetric attacks, Oracle deploys DNS DDoS protection, network DDoS protection, and application DDoS protection.

A Focus on Enterprise Expertise

As stated by Larry Ellison, an Oracle co-founder, and current CTO, “Autonomy is the defining technology of a second-generation cloud.” Oracle has designed its infrastructure to work with many of the most common types of systems and apps. This creates a well-rounded and more user-friendly ecosystem for enterprise customers.

The second generation was made with many different enterprise service needs in mind, including the following:

• Operating systems

• Software vendors for both independent and large corporation software

• System integration

• Digital transformations

• Management of a wide range of service providers within the infrastructure

• Customer support, including IT and integration

• Migration training

• Security 

Knowing that most businesses and IT professionals need some assistance with making the transition, Oracle has developed a set of best practices. 

In early 2020, Oracle announced that it was planning to update its enterprise management platform to further streamline hybrid IT environment development. This demonstrates how the second generation will continue to improve and simplify enterprise manager cloud controls. The Oracle Autonomous Database is perhaps the best example of how the second generation has expanded controls. 

Minimizing Cloud Costs

Initially, migrating to the cloud was not only risky, it was expensive. In addition to thesecond generation streamlining the process, it provides a scalable approach. Oracle has three rate structures:

• Pay as You Go – This allows resources provisioning on-demand without any initial commitment and no minimum service requirements. This means customers only pay for what they use.

• Universal Credits – Customers make a monthly commitment to specific services and receive discounts for those services. This makes it easier to budget the monthly cost and provides the flexibility to change products and regions as needed.

• Bring Your Own License – Customers with Oracle licenses can move them to the cloud, getting the lowest total ownership cost.  

Early versions of the cloud infrastructure tended to lock customers into a single solution. While Oracle made significant changes to its first cloud system, the second generation allows for scalability both up and down based on the chosen pay structure. With improved transparency, it is equally important to be able to scale down services or infrastructure aspects that are no longer required or don’t require as much use.  

A Dedication to Openness

Oracle has always had a significant presence in the open-source realm, and they have continued that with their cloud infrastructure. One of the primary benefits of moving to the cloud is how open and flexible it is, and Oracle values what open source software has to offer. As a member of the Cloud Native Computing Foundation, Oracle wants to ensure companies are able to develop their own standards, such as Helm, Kubernetes, and Terraform. 

Oracle has kits available to help companies develop their cloud space. To give customers the right solution, Oracle’s cloud infrastructure can be tailored for one of three frameworks:

• A public cloud system

• A hybrid cloud system 

• An on-premises system

Open standards are what allow customers to implement their existing tools, scale based on need, and adjust workloads based on the changing needs.