Tuesday, September 04, 2007

The Business Value of Virtual Lab Automation

If you are an ISV, or even if you are managing an internal application development group, you know that application testing and development labs are inefficient with regard to the number of test systems and managing these systems.

Lab systems are usually configured based on the current active development project. Only when a team is actively developing and testing an application is the equipment needed. Given the "peaks and valleys" of development cycles, and given the ease with which even the most complex software configurations can be "spun up and spun down," managing the systems and configurations can be daunting, and leads to either a proliferation of systems, or a huge workload in system configuration / reconfiguration.

This means that a typical development and QA lab requires:
1) Dedicated systems and these systems are woefully underutilized often sitting idle and often preconfigured for support of only a dedicated test set
2) Constant reconfiguration for testing and development support
3) Reconfiguration for replicating client environments for support contact resolution

Besides the direct cost of the inefficiencies, perhaps more important is the lack of lab automation and a better way to manage the configurations can delay development lifecycles, and potentially lead to quality issues (difficult replication of complex issues, less testing because of configuration delays).

The Opportunity

Alinean estimates that virtual lab automation can:

1) Improve net utilization of lab systems by more than 50%, resulting in a typical 60 to 95% reduction in the number of systems required. With a virtualized infrastructure, organization is empowered to create a centralized pool of virtualized servers, storage and networking equipment shared across software development and test teams. With this feature, the team can reduce the number of systems needed to support testing and development lab requirements - resulting in less annual support and maintenance contracts, reduced facilities and space costs, and reduced administrative and support overhead to maintain and manage systems. This helps the team retire or reallocate existing systems, and helps to avoid future growth / system additions.


2) As well, in typical virtualized environments, there is still usually a virtual machine (VM) deployed in the system pool for each physical machine replaced (often even more since it is so easy to create new VMs and few ever seem to get turned off). With intelligent virtual lab automation, it is very easy to turn VMs on and off as needed, reducing the total number that are "live" and occupying space on the systems at any given time. The more teams you have sharing a lab, or the more applications you are supporting, the higher this consolidation benefit will be.

3) With a virtual lab environment, a shared storage library is created whereby users can check out fully provisioned systems "on demand." With this feature, provisioning task time can be reduced by 90% or more through the automatic, rapid set up and tear down of complex, multi-machine software configurations for use in development and test activities. This can help to reduce the tasks and person hours needed to reconfigure systems to support testing requests, helping to improve productivity and reallocate precious resources to more strategic tasks. By empowering the team with self service capability, every developer or test engineer now has the equivalent of their own fully equipped lab / test data center with dedicated provisioning staff.

4) During development cycles, it is often difficult to replicate complex issues when they are found. In a virtual lab environment, the team is empowered to effectively and efficiently suspend and capture “live” multi-machine configurations to a shared library, allowing them to reliably capture, reproduce and share bugs across the team. This reduces the hours spent reproducing bugs for troubleshooting, decreases the constant cycling between developers and testers, and improves overall team collaboration. According to Alinean's estimates, this feature can help to reduce the efforts in setting up environments and reproducing bug / error conditions by 80% or more.

5) Enable automated provisioning and rapid reproduction of software defects, accelerating cycle times and as a result, often increasing the amount of testing that can be performed in a given release time-frame, resulting in a 50% or more improvement in quality (reduction in bugs shipped).

6) Improve time to market responsiveness of the application development group, helping to reduce development lifecycles by 10-15% on average by reducing provisioning / reproduction and support time

The Bottom-Line

With these overall savings, and tallying typical investment in lab automation software (using VMware’s Lab Manager as the benchmark), a typical 100 CPU lab automation project can:

· Achieve an ROI of 1125% on an investment of $105,047
· Derive NPV Savings of $937,000[1]
· Achieve payback in 3 months

Examining lab automation projects in general we find that the benefits are significant:

On average, generates double the ROI of other comparable IT infrastructure projects, 800%+ risk adjusted ROIs vs. 200-400% for most infrastructure projects

Delivers payback on investment of less than 8 months, compared to most IT infrastructure projects which take 10 to 16 months to reach breakeven

Reduces net TCO by >30%, where most infrastructure IT projects deliver only 5% to 15%

Delivers significantly higher availability and agility up-side benefits than projected up-front.

Overall lab automation with virtualization is a low risk, high reward project that any significant ISV or internal application development group should consider today.

For a personalized analysis of the value of virtual lab automation, visit: http://www.vmware.com/go/lm_calculator

No comments: