|
Introduction & Purpose
Data centers experience constant change under pressure from technology and business drivers. IT systems are becoming more, not less complex as new technologies are introduced, and IT staffs are not growing to keep pace with the growth in size and complexity of the typical data center.
Windows™ administrators need to radically simplify both complex and routine tasks if they are going to effectively respond to the constant pressure to deliver increasing value with limited resources. The most effective thing Windows administrators can do to address this issue is to minimize duties associated with maintaining the existing infrastructure.
There are several ways Windows administrators can create greater efficiencies in their daily tasks. This paper describes an automation solution that can significantly simplify many complex and routine tasks, freeing Windows experts to focus on overall application functionality and value to the business. This solution is actually in use today within Microsoft’s own data center.
One Data center Operation that Can Be Radically Simplified
Data center operations are complex largely because of application and infrastructure diversity, and the many locations for storing and using scripts and automation process-knowledge in those environments. Choosing the tasks and functions to simplify through automation must be done carefully and with the right tools. For example, one area of data center operations that can yield large efficiency gains through automation is batch processing.
Almost every IT group uses some form of batch processing. Batch processes are comprised of many tasks and touch a variety of systems as they execute. Most IT groups use multiple schedulers to drive batch processing in their environment, which inherently creates inefficiencies but is better than no solution at all. Therefore batch processing presents Windows experts a great opportunity for efficiency through simplification.
Today, to automate batch processing IT groups use everything from simple scripts and rudimentary scheduling products included in particular point solutions up to more powerful job schedulers. The challenge for many organizations, and therefore the opportunity for efficiency gains, arises out of the fact that each batch processing solution requires its own expertise and each system ends up getting integrated separately and running as a silo, making it impossible for one person to manage the process from end-to-end.
To get a clearer picture of this in the Windows world let us take a look at what standard approaches that most Windows administrators use today.
The Windows Administrator and Batch Processing Automation
Windows administrators will be familiar with these Windows-centric ways to automate and the inherent problems:
- Custom integration by branching from one environment to another, such as calling a SQL script from a batch file using the SQL command line executable, or calling a batch file from a SQL script. This approach uses heavy custom scripting, is difficult to maintain, is difficult to document, and has no clear best-practices.
- Using Windows Task Scheduler. This tool does not currently handle workflows or complex calendars. Even in the “Vista” release it will only runs batch files, executables, or send email
- Scheduling utilities, such as the ones built into SQL Server, MOM, and SMS. These have very limited calendaring abilities, limited job tracking (e.g. status, success, failure) and limited reach in complex, heterogeneous environments.
The example approaches above do not simplify the enterprise-wide orchestration of batch processing. In fact, they offer no consistency: the automation method used often depends on the tool of choice for a particular script-writer at a particular time, rather than a planned approach to automation. Furthermore, they offer no native ability for cross application, operating system, domain, or even network boundaries.
In fact, Windows-centric automation may have little or no ability at all on non-Windows machines. For example, credentials for access to an environment are either stored in clear-text in the script, require user-interaction for automation in that environment, or must be skipped entirely. It is a tradeoff between the security risk of storing clear-text credentials, or depending on humans to run repetitive or iterative tasks.
When steps can be automated, the tasks and the logic surrounding execution such as triggers, conditions, and calendars must be custom coded, often for each individual task.
|
