Cassatt Active Response Basic Concepts: Premium Edition, Data Center Edition

Intended for use with Cassatt Active Response V5.0.

Cassatt Active Response Premium Edition and Data Center Edition automate IT operations by managing shared resources and guaranteeing service levels. This article grounds you in the terminology used in articles on Info Central, and provides a foundation for understanding how you can use Cassatt Active Response.

Shared resources

Cassatt Active Response manages your existing hardware, operating systems, business applications, virtual servers, and network switches (Data Center Edition only) as shared resources. Let's look at the basic shared resources that Cassatt Active Response manages.

In the Cassatt Active Response environment, servers that run your business applications are called application nodes. Cassatt Active Response can automatically discover and manage any application node that is cabled into the Cassatt Active Response network and issues DHCP requests. To fully automate node discovery, allocation, and failure, Cassatt Active Response can leverage a server's power controller (for powering the server on and off) and PXE boot setting for booting the server.

Cassatt Active Response can also manage virtual machines (VMs) as virtual application nodes. Virtual nodes are managed in much the same way as physical application nodes. But instead of cabling hardware, you create a software image that runs your virtual machine manager (VMM)—such as Xen or VMware. The image contains your desired number of VM instances, which you create according to the VMM software you are using. For more information about how Cassatt Active Response works with VMMs, see Understanding Cassatt Active Response Virtualization Manager.

For Cassatt Active Response to manage your business applications, you create a master copy of the software (called a base image), which contains the operating system, application software, and any additional required software. This is done using an interactive command-line tool. From the base image, Cassatt Active Response creates image instances: virtual copies of the base image that it uses when allocating individual nodes. All base images, image instances, and metadata reside in a repository called the image matrix. Cassatt Active Response quickly deploys image instances to application nodes by either booting images over NFS (default), or from local disk.

When an application image needs to be upgraded with new software or a patch, you install the software or patch once, and Cassatt Active Response deploys an updated image throughout the Cassatt Active Response environment without loss of service. And if the patch has a problem, you can easily roll back to previous versions.

Data Center Edition adds the Network Manager, which decouples network configuration from application nodes by introducing programmatic control over network connectivity. Cassatt Active Response can automatically allocate application nodes to business applications that have specific networking requirements—without regard to the servers' current network connections. Cassatt Active Response simply selects a server with sufficient NICs and reprograms its switch port connections to meet the application's networking needs. For more information about how Cassatt Active Response works with networking, see Understanding Cassatt Active Response Network Manager.

Premium Edition and Data Center Edition also support the use of static resources, in which you manually allocate specific application nodes to run each application. Functionality for static resources is limited to the functionality provided in Standard Edition, and is not discussed in this document.

Let's look at the organizing mechanisms for shared resources in Cassatt Active Response.

Tiers

The main organizing mechanism for shared resources is a tier—a temporary collection of resources assigned to do a job. You can think of a tier as a container for:

• An image that contains a set of business functionality.
• User-defined service-level agreements (SLAs) that that enforce how many application nodes are required to provide the business functionality (more on SLAs later).
• User-defined policies that enforce the type of hardware (or VMs) needed for the application nodes, priority of the tier, and whether nodes can be harvested for use in other tiers.
• A set of monitoring collectors that detect node status.
• A set of rules to follow when application nodes fail.
• Allocated application nodes: hardware or VMs assigned to run the tier image.

Pools

Pools contain unallocated application nodes—both physical and virtual. Pools allow you to see what application nodes are available, ready to use, or need attention.

• Discovered – detected nodes, but not inventoried (probed for hardware characteristics)
• Free – available, inventoried nodes ready for use
• Maintenance – failed nodes that need diagnostics or repair

Guaranteed service levels

To automate IT operations and guarantee service levels for shared resources, Cassatt Active Response constantly evaluates the environment to see if it is compliant with the Service Level Agreements (SLAs) and policies that you define. This allows Cassatt Active Response to handle tasks that ordinarily require human intervention, including:

• Detect and replace failed resources and enforce best-fit
• Change resources based on predefined schedules
• Appropriate resources from lower-priority applications
• Reduce power consumption by powering servers off when not in use (and back on when needed) using Active Power Management technology

Rules engine: optimizing flux by enforcing policies and SLAs

The rules engine optimizes flux by the policies and SLAs you define. Policies tell the rules engine how to react in certain conditions. For example, for a node failure, your policy might be to reboot the node, then, if that doesn’t work, to move it to the maintenance pool and get a replacement from the free pool. SLAs tell the rules engine how to enforce the number of application nodes required to provide the business functionality.

Policies and SLAs are configured on tiers and roll up into a group of settings, as shown in the following table. No scripting is required to configure Cassatt Active Response for full automation.

Active Power Management technology

Active Power Management technology introduces editable power policies based on schedules or on-demand events:

• Scheduled – you can schedule Cassatt Active Response to execute different power policies to accommodate off-peak hours or specific days when machines are not in use.
• On-demand – Cassatt Active Response can monitor external events and execute policies that decrease power based on availability or cost; for example, power incentive programs like Pacific Gas & Electric (PG&E) provide rebates to companies that can meet negotiated power reductions. Alternatively, you can choose to invoke a policy "Now" from within the Active Response UI.

For example, you might create a power reduction policy for Sundays that reduces the tier targets when machines are not in use. You might create another power reduction policy that reduces power by 50%—to achieve a prenegotiated reduction with a utility company (for example, PG&E).

A power policy can have two components:

• Tier settings –define the service levels for each tier when the policy is invoked: on/off, or running at a specified operational target.
• Schedule (optional) – defines the day(s)/time(s) when the policy is automatically invoked. Each policy can have multiple schedules.

Whether or not you define a schedule for a policy, you can execute policies on the fly in the Cassatt Active Response UI.

Metering and monitoring

Cassatt Active Response collects and displays a range of monitoring data about the status of node hardware and software running in the environment, allowing you to modify SLAs and policies as trends and patterns inform your IT operations. For custom reporting needs, Cassatt Active Response normalizes and stores monitoring data in an internal database that is accessible to external monitoring applications.

Components

Cassatt Active Response runs on a single server, called a "control node." The control node hosts three different web-based user interfaces that you'll use for different purposes, as shown in the next illustration. Optional Cassatt software for special uses is available in feature packs.

• Controller – A web-based application for configuring SLAs and policies and for viewing the Collage rules engine in action using dashboards. This is where you'll arrange your nodes into tiers and specify tier requirements.
• Policy Manager – You'll use this UI to configure power reduction policies and their associated schedules, and to activate policies on the fly. Tier managers can change participation in policies for their own tiers, as well as set tier operational target values.
• Report Manager – The reporting UI is designed to give executives and data center personnel a month-by-month view into application uptime, estimated power savings, and more, with summary charts and detailed listings of what was on and what was off and for how long.
• Utility company feature packs – If your company is participating with a utility company rebate or power reduction program, Cassatt may have software that can accept power reduction policy activations directly from your power company. See Feature Packs, or check with support@cassatt.com for more information.
• Automation feature packs – Casssatt provides special-purpose software for enhancing automation in common environments, such as web services. See Feature Packs, or check with support@cassatt.com for more information.

top

Summary

The Cassatt Active Response software, with its sophisticated rules engine, manages shared resources and guarantees service levels, which optimizes IT operations. Cassatt Active Response software is installed on servers called control nodes.

Shared resources comprise application nodes (physical and virtual), software images and their image instances and, with Data Center Edition, network switches. Two important organizing mechanisms are tiers and pools; tiers contain temporary resources to do a job (application nodes, image instances, SLAs, policies), and pools contain unallocated application nodes.

Ready for a little more depth? I recommend the companion article in this series: Sample implementation: Premium Edition, Data Center Edition. This article goes through each step in a basic Cassatt Active Response implementation—from cabling hardware, to getting Cassatt Active Response up and running. If you are preparing to start an implementation, or just want to know the tasks involved, this is a must read.

Was this article useful? Tell us what you think.
Email infocentral@cassatt.com.