1 Cloud Management - An Overview

Planning

Using Enterprise Manager, you can setup a Cloud with brand new hardware, new software, and even a new data center. It allows you to transform existing data centers into a Cloud environment. Before setting up a Cloud, you need to plan the infrastructure requirements such as the physical and virtual networks, storage arrays, applications and so on. The auto discovery feature in Enterprise Manager helps discover and baseline all physical and virtual IT assets in an enterprise.

The Enterprise Manager Consolidation Planner is a powerful tool that helps administrators plan the Cloud architecture. It allows administrators to identify source and destination targets and applicable technical and functional constraints such as where the application can reside, and so on. Administrators can generate consolidation advisories that may include plans to move from Physical to Virtual (P2V), Physical to Physical (P2P), or Physical to an Exadata solution. The Consolidation Planner can also be used to identify the database consolidation plan which is helpful when setting up Database-as-a-Service (DBaaS).

Setting Up

Enterprise Manager can be used to model infrastructure-as-a-service (IaaS), middleware-as-a-service (MWaaS), and database-as-a-service (DBaaS) clouds. It offers capabilities such as bare metal provisioning of Hypervisor, setting up of server, storage pools and grouping them into zones based on functional or QoS characteristics. Enterprise Manager leverages the Virtualization Storage Connect technology, where the Cloud setup process is integrated with storage technologies like Netapp, Hitachi, Fujitsu.Administrators can define standardized service templates for databases and middleware platforms, and publish these as services. These services can represent single-tier templates or complex, multi-tier enterprise platforms. Enterprise Manager uses the Oracle Virtual Assembly Builder (OVAB) to help package a multi-tier platform into a single metadata driven Cloud service. Using OVAB, platform architects can model the entire platform topology graphically, define all dependencies, deployment constraints, and deliver the entire stack in the form of an assembly.

This assembly can then be published to the centralized Software Library in Enterprise Manager, and be made available to developers as a Cloud service – an entire application development stack, that can be provisioned quickly. Administrators can create different types of services depending upon the business needs. For example, administrators may offer a database service based on different versions of the Oracle database, but only the ones approved for use within the business. Enterprise Manager comes with a role-driven access control. Integration with LDAP allows Enterprise Manager to inherit enterprise roles. The resource limits are implemented with quotas that are tailored for the specific service type (IaaS, MWaaS, and DBaaS). This prevents unauthorized usage of a service while also preventing a few users from using majority of the resources in the Cloud.

Building

Enterprise Manager allows entire applications or components to be packaged and published to the Cloud as a service. This expedites application development and provisioning processes within an organization. Developers can publish utility components and applications in the form of assemblies and service templates for reuse within their groups. Similarly, allowing applications to be available as assemblies allows testing teams, business analysts or production teams to deploy pre-built applications in a few clicks.

Testing and Deploying

After an application has been built, it needs to be tested. Enterprise Manager provides a testing portfolio that allows users to test both application changes and changes to the database. The testing solution provides the ability to capture a production load and replay in a test environment, so that the results are predictable. The testing solution also leverages the diagnostic capabilities built into the technology layers and provides prescriptions for remediation.Enterprise Manager provides a self-service application that lets end-users deploy a service. This self service application can also be customized. End users can choose to provision application assemblies, along with databases and platforms, in an on-demand manner. For each request, they can specify the amount of underlying resources such as CPU, memory, and so on that they require for each component. Enterprise Manager automatically provisions the requested service and the appropriate resources. The self-service application also lets users define policies to scale out or scale back resources based on schedule or performance metrics. For example, one can set a policy to elastically scale out a Web Server if the processor load on existing Web Servers exceeds a certain threshold value.

Monitoring and Managing

Enterprise Manager provides the ability to collate targets into groups for better manageability. The Administration Group feature allows administrators to define monitoring settings, compliance standards and cloud policies through templates and also organize each target in multiple hierarchies, such as Line of Business and Lifecycle status. This allows the monitoring framework to scale to thousands of servers, databases and Middleware targets in the Cloud. Enterprise Manager comes with an in-built Incident Management system that can manage by exceptions. Administrators can review, suppress, escalate and remediate the events as needed, and also integrate it with ticketing systems.Enterprise Manager has the ability to define contractual Service Level Agreements (SLAs) that govern the contract between the application owner and the provider of the Cloud. Administrators as well as users can also define management policies that automatically adjust the service resources to ensure that SLAs are met.

Enterprise Manager also provides user experience management and business transaction management. The Enterprise Manager Configuration Management capabilities are optimized for Cloud environments. It can monitor vast numbers of configurations continuously, discover changes, measure drifts, pin-point configuration errors, as well as offer insight into system topologies, within a single Console. Enterprise Manager Cloud management capabilities are also integrated with My Oracle Support. This integration delivers facilities such as Patch Advisories, Service Request Management, Knowledge Management right on-premise and in-context of the overall Cloud.

The IaaS and DBaaS Home pages in Enterprise Manager allow Cloud administrators to get a summary view of the requests, the general state of the service such as zones, pools, servers, and databases. The IaaS Home page also provides a peek into popular policies and summary of incidents.

Metering, Charging, and Optimization

The Metering and Chargeback features in Enterprise Manager support basic metrics like CPU, memory, and storage usage, and offer pricing models based on application usage, database usage, and Middleware-level metrics. Administrators can also extend the pricing models to account for fixed costs, configurations, administrative expenses, people costs, energy utilization or a combination of these.These capabilities enable enterprises to account for actual usage versus representative usage. Cloud Management also entails an ongoing optimization of resources as well processes to make sure that the service levels are persistent. Enterprise Manager provides administrators and application users with features that help rediscover assets, re-evaluate the performance, rebalance the Cloud, and fine tune the provisioning process. The tuning capabilities in the operating system, database and middleware layers aid in continuous optimization and subsequent improvement.

IaaS Cloud Anatomy

The IaaS Cloud Anatomy consists of the following:

• Server: A physical computing system. Typically, a server runs a hypervisor and is part of a server pool.

• Server Pool: A server pool is a set of tightly coupled group of servers (typically up to 32 servers) that hosts a set of Guest VMs. The servers are assumed to be largely homogeneous in capabilities and in connectivity. High Availability and Live Migration is permitted within a server pool boundary. Each server pool needs to have access to a shared storage subsystem (which could be just be a NFS mount point) to facilitate live migration. In addition, access to a clustered file system may be required to maintain the HA heartbeat file.

• Storage Entity: An individual file system or block store. Each storage entity is served by a storage pool. Some entities are free standing and will exist until they are deleted. Other storage entities that are associated with one or more Guest VMs are deleted when those VMs are retired.

• Storage Pool: An abstract storage system that hosts a set of storage entities. A storage pool is generally implemented using a set of storage devices such as disks, SSDs, and storage servers.

• Zone: A Cloud could consist of one of more zones. A zone is a logical grouping of resources (for example, servers and storage) to facilitate self-service provisioning and administrative purposes and it could typically consist of hundreds to thousands of servers. A cloud consumer (self-service user) could, for instance, specify that his request needs to be provisioned within a given zone.

  Zones are non-overlapping - meaning that a resource can only belong to one zone. However, resources within a zone may be accessible from another zone. For example, a VM in Zone 1 could interact with a VM in another zone. A zone consists of a set of server pools or just a loose grouping of individual machines. The second case may be simple to set up and will not require shared storage; however no HA and live migration is permitted within this zone.

• Cloud: A set of storage pools, server pools and zones under the programmatic control of a Cloud Controller and the administrative control of the Cloud Administrator. The Cloud Administrator works with the Cost Center Administrator who has paid for the cloud to determine a resource allocation and charge back policy that meets their needs. A very large company or a company with complex security requirements may have more than one cloud.

DBaaS and MWaaS Cloud Anatomy

The DBaaS and MWaaS Cloud Anatomy consist of the following:

• PaaS Infrastructure Zone: A PaaS Infrastructure Zone is a group of homogeneous resources such as hosts, or other targets. Each resource in a zone represents a location at which a service instance is to be deployed.

  Before you enable or setup DBaaS or MWaaS, you must create a PaaS Infrastructure Zone which allows you to define the placement policy constraints for a specified set of targets and the users to whom this zone will be available.

• Software Pools: For DBaaS, you need to set up Database Pools. A Database Pool is a collection of database homes. For MWaaS, you need to set up Middleware Pools. A Middleware Pool is a collection of middleware homes. A Database or Middleware Pool has the following constraints:

  • A target can belong to only one software pool.

  • The name of the software pool and the version cannot be modified after it has been created.

  • All targets in a software pool must be homogeneous.

  • A Service Template can use multiple zones but only one software pool within each zone.