IBM Workload Deployer – Continuing Maturation of Private Cloud Appliances
IBM Workload Deployer is the next generation of the IBM WebSphere CloudBurst v2.0 appliance further driving the use of patterns, templates and automation to simplify and accelerate the definition and deployment of private cloud solutions. The change in name signifies a growth in the scope of the appliance beyond just WebSphere components. With Workload Deployer, IBM is providing Hypervisor Edition Images of WebSphere Portal Server & Lotus Web Content Management, DB2, WebSphere Process Server, WebSphere MQ, WebSphere Message Broker, IBM HTTP Server and WebSphere Application Server. In addition to having a broader scope, Workload Deployer has more processing power, storage and network capacity than the WebSphere CloudBurst appliance.
With a private cloud, there needs to be support for creating standardized images, virtualization and automation. Workload Deployer supports deployment customization and standardization through a selection of predefined components, patterns, policies, cross-component links and templates.
At launch a pattern provided by IBM is the IBM Workload Deployer Pattern for Web Applications. This pattern provides a PaaS solution to which an application packaged as an EAR or a WAR can be deployed. The workload pattern encapsulates the installation, configuration and integration of middleware components. The pattern user is then able to focus on the deployment of their application. Topology patterns are also available and provide more flexibility and configurability in regards to component installation, configuration, and integration.
In addition to the content supplied with Workload Deployer, it is possible to use custom built Hypervisor Edition Images, patterns or templates. The IBM Image Construction and Composition Tool can be used to create custom built Hypervisor Edition Images.
Workload Deployer also provides support for a number of virtualization platforms including VMware ESX (Linux), PowerVM (AIX) and z/VM (Linux on System z). In addition, there is support for Autonomic Elasticity – allowing virtual systems patterns to scale up or down based on observed activity, and there is support for VM mobility to adjust VM placement to optimize resources.
Workload Deployer shows great promise by delivering a balance between standardization and automation of an appliance and customizability and configurability needed to meet the needs of an organization. However, content in the form of Hypervision Edition Images, patterns, and templates will be the deciding factor in the success of this approach. At a minimum, this will put pressure on other vendors to create private cloud tooling to work toward a similar balance and deliver content for their platforms.
Some nice blog posts with video overviews of IBM Workload Deployer
I created links to them in my recent blog entry:
Check out the basic IBM Workload Deployer overview video:
I'd also suggest you check out how simple concepts like elasticity are given an application focus with policies (as opposed to a server and dependent technology focus). I believe this is really revolutionary in the platform as a service space:
This approach does require the user to release some control from their application design (for example, how is the elastic scaling achieved with regard to stateful session data), but that is likely easy for many of the applications in the real world -- they just care that it works, not how it works.
I've seen other vendors talking about middleware as a service or cloud technologies, but many times they don't come with this level of self service and middleware as a service abstraction. Many times they are just infrastructure as a service with hand assembled middleware.
Note that I work on IBM Workload Deployer and associated technologies, so I can explain much of this if asked. My main job is to make this technology highly performant. You can see some of this work in action in the following video:
I have to say I'm pretty excited to be showcase self-service provisioning of elastic and highly available configurations of application servers and databases in minutes!
Tom Gilb & Kai Gilb Jan 26, 2015