Storage automation Part 1: Pot of gold or can of worms?

Computerworld - IT organizations are looking to automation as a way to operate more efficiently, but they are also grappling with two burning questions: Do automation tools simplify or complicate daily tasks? And do the rewards of automation outweigh the effort and risks?
When I started my career in the late 1960s, one of my first assignments was to design automated processes for testing complex equipment. For the most part, this consisted of taking the giant three-ring binders that contained thousands of step-by-step instructions and checkpoints for every conceivable error condition and then designing special-purpose equipment to perform a small part of the process.
Nearly 40 years later, I am back working on tools to build automated processes, this time for storage, and I'm doing it with software. In many respects, the issues and challenges we face today in enterprise storage are the same as they were back when I started out in the defense industry. We still need to define what steps to follow, what inputs are required, the proper order of the processes and the error conditions that need to be handled.
We now have more powerful tools to work with, but the tasks that we face are substantially more complex. Given a limited number of skilled people and sometimes conflicting requirements, IT organizations need to understand whether the value of automation is worth the effort and if the rewards outweigh the risks.
Provisioning complexity
From discussions with storage professionals, one of the areas most often identified as a choke point is provisioning of storage within a storage-area network. Provisioning is certainly among the most critical and complex tasks that storage administrators face on a regular basis, and not surprisingly, one that can cause great harm if done incorrectly. To put this in perspective, in talking to enterprise customers, we have identified more than 50 discrete steps as necessary to complete provisioning of additional storage resources for an application such as Exchange.
Typically, one of the senior storage administrators is called upon to expand storage capacity following a process based on his individual experience and expertise. Given the number of steps and the varied backgrounds of multiple administrators in an organization, there is a nearly infinite number of paths that can be followed. Most of them are correct, but each one is unique.
In isolation, there is nothing wrong with any of the varied choices made, but over time, these different processes introduce a level of complexity that can result in chaos when a problem arises. Something as simple as a difference in naming conventions can have a ripple effect down the road. How can we bring some sense to such a complex problem?
Many of the major hardware vendors provide best practices for provisioning storage on their devices, and some even have tools to help with the process. While useful, these offerings only address a small part of the job; beyond hardware and file systems, we must address security, data protection, networks and the applications themselves.
Given the criticality of the process, some level of management approval should also be incorporated. What's needed is an overall supervised process, driven by best practices that address the complete provisioning cycle from end to end. The process must include initial verification of the need for additional capacity, assigning space, configuring logical unit numbers, expanding volumes and dozens of other steps.
Once the additional storage is in place, there must be final reporting on the new configuration. Related tasks, such as setting up a backup policy, modifying security settings and updating documentation, must also be established. Only when every step is set down and enforced will the goal of consistency in operations be realized. Most of the same issues and challenges facing storage professionals in devising a process for storage provisioning apply to many complex, as well as routine, tasks in storage management. Let's look at the broader challenge of automating storage processes in general.
What do we mean by automation?
When I hear the word automation, I immediately think of autonomous, lights-out operations that run themselves without human intervention. But taking humans completely out of the loop is rarely practical and can have negative consequences. Remember HAL, the computer that controlled nearly everything on a spaceship in the 1968 Stanley Kubrick film 2001: A Space Odyssey?
Completely autonomous operation is certainly a worthwhile goal, but it won't be practical for complex storage management processes in the foreseeable future. So when we talk about automation of storage processes, just what do we mean? Here are the critical components as I see them:

• Encapsulation of best practices into repeatable processes
  
• Supervised execution of the discrete steps through a simple-to-use interface
  
• Automatic requests for approval of specific actions based on policy
  
• Monitoring and management of the execution of processes
  
• Integration with enterprise storage management tools providing seamless operation
  
• A complete audit trail showing who did what and when they did it
  
• The ability to change individual actions from supervised to autonomous, when they become trusted processes

Read more about storage in Computerworld's Storage Knowledge Center.