1. The Context
Virtualization in Pre-Production Labs: Virtualization usually takes its first steps in an organization in the pre-production (software lab) environment. The use of virtualization in dev/test labs continues to be a very popular use case today. I would think that it would still figure in the top three use cases for virtualization, much like how this earlier survey in 2007 indicated. As IT administrators get familiar with the use of virtualization, the technology moves towards being used in the data center environment as well.
Virtualization in Production Data Centers: The big 4 infrastructure management vendors – BMC, IBM, CA and HP have recognized the impact of virtualization in the data center and have been taking initiatives towards managing a hybrid of virtual and physical infrastructure. Alessandro recently suggested that VMware would join the list and become an infrastructure management company as well. Virtualization adoption in the data center is an important development and most IT practitioners believe this to be a fundamental shift in the data center operational dynamics (think of an agile and dynamic data center) that is here to stay.
Hypervisor Commoditization: The adoption of virtualization in the above two pre-production and production arenas has been accelerated by recent industry announcements. With the base hypervisor technology moving towards commoditization – the value value add comes from the accompanying management applications.
2. Contrasting the Management Environments in Pre-Production (Lab) and Production (Data Center) Environments
Given the above widely accepted view that virtualization has significant impact on IT infrastructure operations in the pre-production (lab) and production environments – how do the accompanying management applications in each arena compare? The figure below contrasts the management environments in the two cases.
As the table indicates (see the management goals specifically), there is a fundamental difference between the two environments and therefore there is a need for separate management technologies with different management functionality in the lab and production environments.
3. Specific Benefits of Virtual Lab Automation and Management Applications in Pre-Production (Lab) Scenarios
Pre-production (lab) scenarios bring with it their own set of management challenges which are not addressed by production management applications. Here is a snapshot of the management challenges in the lab for which virtual lab management solutions provide a good solution (and where production management apps fall short):
|Lab Management Dimension||Why it is needed||Product Demonstrations
|Web based self service||Allow multiple users and teams to share and leverage a common set of shared infrastructure on demand||Enabling web based self service|
|User management||Support a large internal constituency of users and teams that need access to lab infrastructure||How User Management Helps in Test Lab Operations|
|Multi-machine management||Ability to create, deploy, share and delete multi-machine configurations as a single unit||Automating the Creation of a Multi-Machine Test Environment|
|Rapid changes to VMs and Configurations||Easily allow users to create and edit multi machine configurations and VMs||Automating the Creation of a Multi-Machine Test Environment
|Sharing and Collaboration||Allow the many users to share lab artefacts and leverage work done by others (without the need to re-do the effort)||How User Management Helps in Test Lab Operations|
|Network zoning||Ability to isolate running configurations so they do not conflict with the same config instance deployed multiple times||The Benefit of Network (IP) Zoning in Executing Test Environments|
|Storage optimizations||Optimize the storage occupied by VMs, especially when hundreds of VMs are created by the large internal lab user community||The Benefit of Linked Clones in a Dev/Test Environment|
|Prevent resource hogging||Setup policies and quotas for individual users so their usage of the lab remains bounded. Also setup job lease times to prevent a single user’s job occupying all the server capacity for prolonged durations.|