Automating reviews and checkpoints

One of the more sensitive and critical issues in following best-practice guidelines is the issue of reviews and approvals.

In most best-practice recommendations, there are checkpoints when formal reviews or approvals are indicated along with implied or explicit recommendations for documentation and sometimes even specific forms that need to be filled out. These checkpoints are typically well-thought-out process or project guidelines, and the approvals can often mean the difference between consistent and well-governed efforts and random, ad-hoc responses to problems and requests as they crop up.

However, formal, manual reviews take time and can seem to “slow up the process.” Sometimes, in fact, best practices can be implemented so rigidly that they lengthen routine responses to service and infrastructure performance problems.

There is no simple answer for this, except to say that all best-practice initiatives should be viewed as journeys rather than destinations. Their single greatest value is to stimulate you to think more deeply about processes, requirements and team interaction. Understand the recommendations and apply them thoughtfully to you organization.

There is another approach to these “checkpoints,” however. You can balance the need for accountability and validation with the need to be responsive and time-sensitive by buying management products that embed validation and accountability in an automated fashion, so that reviews can be provided in the way of reports with clear visuals about actions taken, when they were taken, where and why. In other words, the checkpoints become automated actions within your management products, with or without a human being passing judgment on them in real-time.

There is potentially risk associated here, but with the right software for the right processes, operational savings – and sometimes even gains in accuracy and effectiveness – can be huge.

Here are places where software can help.

Examples in change and configuration management software include AlterPoint, Emprisa, Intelliden, Network Clarity, Opsware and Voyence on the network side and Altiris, BladeLogic, Configuresoft, Ecora LANDesk and Opsware (TripWire) on the systems side (BMC, Computer Associates and HP also have invested in systems change management). These vendors’ products are designed to automate and standardize changes to the infrastructure. They automatically generate audits and in some cases automatically generate compliance reports. Most of these have fairly sophisticated ways for embedding policies and standards, so that any deviations are immediately made visible. Minimizing and focusing human involvement should actually minimize errors. Manual processes, even manual reviews, may be moments of error in change and configuration management, as different teams approach device configuration in support of different application services and sometimes make conflicting changes to the same device.

Performance and availability management – or real-time service assurance (“Incident Management” in the IT Infrastructure Library’s processes) – is another area for advanced, highly automated, dependable and time-sensitive management software. This market is so vast that doing it justice would require at least a feature-length article, and possibly a book. Suffice it to say that trouble-ticketing systems lend themselves to formal checkpoints, individual actions and reviews, but many of the more advanced management products that do root-cause analysis, application flow management, service-level management and application acceleration provide levels of automation that outpace traditional help desk approaches, sometimes with greater accuracy.

There are also products designed to validate that a fix has been made correctly. Enterprise Management Associates research has documented what you probably already know: most IT organizations function have lots of manual approaches and expert opinion with little or no validation. (In many shops, validation that a fix has been done correctly is when the phone stops ringing.) The right mix of management software here can vastly improve these processes if deployed well. Also, management technologies are evolving to become more “deconstructable” in themselves – so you can see exactly what was indicated, where, why and when and in what context. This is essential if you’re going to rely on management software to automate and to some degree account for processes and actions.

These are just examples of areas where “reviews/checkpoints” can be maintained in sympathy with best-practice guidelines, but partially or even largely automated through good management technologies. Both are areas where time sensitivity is high and where processes are routine and ongoing. Planning major initiatives, whether for introducing a new application service or doing data center consolidation, have a different scale and time dependency.

In all cases, however, best-practice guidelines remain invaluable and fundamental in clarifying and mapping out process requirements. What changes is just how manual all “reviews” remain. In the Internet era, when some services need to be provisioned, delivered, assured, accounted for and terminated within minutes, a balance between well-chosen technologies and formal oversight provides a healthy recipe for success.