IBM Cloud Manager with OpenStack lets developers build common cloud workload task automation into an application -- Kane Scarlett, developerWorks
DevX.com editor Jason Bloomberg once wrote a good definition on cloud workloads: "The best way to think about a Cloud workload is [as] all the individual capabilities and units of work that make up a discrete application." I'd like to take that thought further and say a key component to successful cloud computing is
the ability to service varied workload requests ...
A data-analysis application would require a differently configured workload than a simple communication-oriented app, and so on. The service tasks I'm talking about would have to include performing the input analysis needed to determine the changes to make and the resources to use and executing those decisions.
By "dynamically," I mean "real-time and automated."
Some of the most cloud-friendly types of workloads are those that require
Of these three, only number two is a task you might contemplate handling manually. Even then, automation can ease the setup and processing of this sort of task by managing a potentially complex schedule of processor resource timing.
It is coming to the point in which the entire span of IT infrastructure is controlled by software; the Software Defined Environment (SDE) relies on automation and integrated models of expertise and experience to exist. And application developers can no longer be content with just the concerns of application development -- as a habitual practice, they should also consider the workloads that support their applications.
That area seems like a perfect place for some expert automated assistance. Then the developer can focus most of his effort on what is most important to him -- the development of an award-winning application.
Why developers should be familiar with IBM Cloud Manager with OpenStack
As companies increasingly migrate to SDEs, automating hybrid cloud infrastructure across multiple platforms through open technologies becomes essential. Open means it's easier for you to adopt a cloud model and integrate it with your existing applications.
You probably know this product as SmartCloud Entry, but the latest incarnation of it is called IBM Cloud Manager with OpenStack. Regardless of what it's called, it is solid cloud management software based on OpenStack that is integrated with IBM enhancements and support. It provides support for the latest OpenStack operating system release, Icehouse, and full access to the complete core OpenStack API set to help clients ensure application portability and avoid vendor lock-in. It also extends cloud management support to System z, in addition to Power Systems, PureFlex/Flex Systems, System x, or any other x86 environment. The new solution also provides support for IBM z/VM on System z and PowerVC for PowerVM on Power Systems designed to add more scalability and security to Linux environments.
IBM Cloud Manager with OpenStack is a self-service portal designed to simplify cloud management for the cloud user. It enables you to work with virtual appliances and workloads focusing on the user’s perspective, delivering such self-service capabilities as provisioning and de-provisioning servers, drafting and cloning deployments, taking deployment snapshots, starting up and shutting down servers, and resizing existing servers.
I mentioned two cloud-friendly workloads; one that required unpredictable scaling and one that required additional CPU processing. In this demo, you can see how IBM Cloud Manager with OpenStack automates scaling for a workload of this type, one in which the presenter has overtaxed the system with the need for about 50 percent more CPU processing resources than is currently available. It takes a bare-metal system, provisions an operation system to it, and then expands the GPFS shared storage file system onto the node. It also deploys OpenStack onto the node and expands the OpenStack cloud. At the end of the demo, the visual display has changed -- the system had scaled up to an additional hypervisor and has rebalanced the virtual CPU utilization; it has also automatically added the additional needed storage. (Another video provides a more admin point of view of the process; it demonstrates how to use the product to deploy a ready-to-go OpenStack-based cloud from a cluster of bare metal servers.)
One of the powerhouse pieces of Cloud Manager is IBM Platform Resource Scheduler; this software product delivers enterprise-class dynamic resource management for OpenStack cloud environments that enables:
The previous demo includes the roles that PRS plays in the system.
IBM is offering a beta program for Cloud Manager that lets your on-premise cloud reach out to a public SoftLayer cloud for resources during utilization spikes. As I get an opportunity to speak with the designers of this feature and understand the mechanics of how it works, I'll report back to you on this feature.
Try IBM Cloud Manager with OpenStack
Download the trial version (or hosted trial) of IBM Cloud Manager with OpenStack, including IBM Platform Resource Scheduler and try all of the functions of the product for no charge, for up to 90 days. Cloud Manager with OpenStack trial code is available for IBM Power Systems, IBM PureFlex and Flex Systems, and x86 servers.