UniCloud Live Demo Wrap Up: How To Create A Hybrid Cloud Computing Environment

    This post was written and contributed by Gary Tyreman, SVP of Product & Alliances of Univa, a Rackspace Cloud Tools Partner.

    Thank you to Rackspace and your many users who attended our webcast and demo last week on hybrid cloud computing. We had a great time demonstrating how easy it is to virtualize computing premises using Univa to, in this case, extend your internal computing infrastructure to The Rackspace Cloud.

    I began the webcast with an introduction to Univa’s view of the next generation data center, where “On-premise” is a relative term. The reality is that large organizations and service providers need to get much more efficient to be able to compete – they can’t be constrained by physical or political boundaries that keep resources from being used to their maximum efficiency level. In the next generation data center you need to be able to run any application anywhere (internally or externally) without worrying about compliance constraints, data requirements, costs, or performance.

    I then discussed how Univa makes this possible with our cloud management solutions. Our approach uses ‘premise virtualization’ to provide common management for the control and federation of multiple hybrid clouds. Basically, premise virtualization takes server virtualization plus a form of application virtualization and extends this environment with provisioning virtualization. The result is a new sense of location independence, where applications run on the optimal available resource. I went into some detail about how we make this possible via 4 key solution factors:

    • Application Virtualization
      • Enables a new sense of location independence
      • Encapsulates organizational know-how
    • Premise-Specific Systems Provisioning
      • Common management of any resource in any premise
    • Automation
      • Enables wider use of higher order policy management
    • Control Policies
      • Control, reducing risk
      • Policy management: What goes where, if, when
      • Rapidly exploit price and functionality fluctuations in any cloud

    I talked about how Univa does this by packaging the application, and everything it needs to run, into ‘Kits’ which combine with native packages to assemble the software stack on a Software Profile. This combined with the Hardware Profile is shipped to the computing premise – in-house or public, physical or virtual – depending on where it makes the most “sense” for the job to run based on what the user cares about (as encapsulated by policies, requirements, compliance constraints, or economics).

    Next, my colleague Bill Bryce showed participants a live demo of our UniCloud product creating a hybrid compute environment extended from his laptop to The Rackspace Cloud. Bill showed how you can burst capacity into Rackspace easily and can track jobs from submission to completion and node tear-down via our management GUI.

    Here are some highlights from the Q&A session:

    Q: Describe how we set up and use policies.
    A: The Univa policy engine works on either a polling or event basis.  It reviews rules based on where metrics data is coming from and determines what action to call based on that data.  For example, the system can look at the number of nodes running in one cloud provider and, if this is too many, move nodes to another provider. Or, one can configure the system to (for example) always try to provision VMs internally before moving workload to a public cloud. In short, any rule that can be dreamed up can be put into engine.

    Q: Does Univa provide a billing infrastructure?
    A: What Univa does is consolidate bills so the customer gets only a bill from Univa rather than Univa plus a number of cloud providers. This makes things easier for the end user. We can also work with the enterprise to bill based on departmental use.

    Q: How does this work with central application licensing?
    A: The nodes in (for example) Rackspace’s cloud can access the license server in the private internal environment – should show up as nodes with IPs in the network environment. [Bill showed this in the demo GUI.] Essentially, the Rackspace compute infrastructure is extended into the on-premise environment – Univa makes it look like they are part of your network.

    Q: Where is the software available?
    A: The base framework software can be downloaded at http://www.unicluster.org/download. The value added kits that allow you to create a hybrid cloud (such as the Rackspace adapter) can be acquired from Univa – contact us at http://www.univaud.com/about/contact/request-info.php

    Click below to watch the entire webinar.

    We’d be happy to set up a private demo for you – contact us at http://www.univaud.com/about/contact/request-info.php.

    Rack Blogger is our catchall blog byline, subbed in when a Racker author moves on, or used when we publish a guest post. You can email Rack Blogger at blog@rackspace.com.


    Please enter your comment!
    Please enter your name here