April 21, 2014

Interactive Analysis Tools (calculators)

Keep in mind the following caveats before you begin using the high-availability decision tree:
  • This tool is not meant to be precise. Determining an application's tolerance for failure can be a subjective decision and can change due to different circumstances. The conditions and factors here have been simplified for general usage.
  • While the following six questions reflect the issues that will primarily determine an application's need for high-availability clustering, other unspecified challenges may tilt the decision in one direction or another. The more diverse a group of applications and operating systems included in a high-availability environment, the more complicated the decision becomes. Complexity increases risks in a geometric, not a linear, way.
  • The values in this tool are based on current practice. Right now, high-availability costs are subject to two opposing trends: the cost of hardware and software is going down, and the cost of people is going up. However, improved systems management, such as that provided in the Microsoft® Windows® 2000 Advanced Server edition, is helping to reduce the people costs. Future high-availability costs may not reflect the values and issues presented here, as tools and techniques continue to evolve.
  • Legal disclaimer: Microsoft® Corp. provides this Rule of Thumb analysis tool solely for self-evaluation. The tool is offered "as is." Microsoft, applications vendors, network services vendors and independent consultants are the primary sources of useful information and analysis about clustering as a tool to enable 24x7x365 reliability.
If you have any questions, comments or suggestions about this Rule of Thumb tool, please contact the editor of the Enterprise from Microsoft at: entfeed@microsoft.com. Unfortunately, the editors are not equipped to respond to questions about Microsoft software or other topics unrelated to the material on the Enterprise Web site. For technical support, please go to http://support.microsoft.com/directory/

The following tool takes you through six questions that will largely determine if high-availability clustering makes sense for a given application. The values presented here were developed based on interviews with analysts and IT managers who have implemented high-availability clusters, as well as other research.

To use the tool, simply select the numeric factor for each item that most closely reflects the application under consideration.

Question 1
How long can the organization tolerate the application being unavailable?

99% uptime (3.65 days per year down)

99.9% uptime (8 hours, 45 minutes per year down)

99.999% uptime (5 minutes per year down)

Greater than 99% uptime (Availability not a concern)

Question 2
To estimate the cost of downtime on the application being considered for clustering, enter the following information:
(Please use numbers only in the blanks below. No letters or symbols, such as $.)

A. Number of workers using the application:


Average hourly cost, including benefits and other allocations:

B. Number of transactions per hour:


Average value of each transaction:

Interested in learning more about our portfolio or have a specific custom content project? Contact Larry Marion at: lmarion@triangle-publishing.com