IT Focus Area: Connectivity
July 28, 2012
7 Things to Consider When Automating the Mundane
A decision-making framework for reviewing automation solutions will help you see how they align with the priorities of the business. There are seven areas to consider when automating the mundane.
1. The type of automation engine
Some engines are meant only to assist engineers in doing specific tasks, whereas others are designed to dramatically reduce the labor required for incident resolution. The needs of the business and the makeup of the IT operations workload will determine which type of automation is required. Only the new systems with expert system cores can reduce day-to-day operational effort. But if your highest priority is provisioning, traditional orchestrators are adequate. However, note that the price differential for many customers between standard automation and expert systems is eroding.
2. Technology lifecycle
Some automation engines are focused on a specific phase of the target technology lifecycle. For example, most of the server and virtual machine automation covers only the provisioning of new servers or retiring of existing ones, but offers little in day-to-day operations except for alerting. Expert systems, on the other hand, can cover most of the phases in the lifecycle.
3. Technology scope
Most automation platforms are built to address a specific class of devices or applications. Specialization simplifies the building and deployment of operations systems. On the other hand, too many specialized systems means a great many of them have to be deployed to cover all the needs of an IT organization. They are notoriously hard to integrate, which places serious limits on the productivity of the IT staff overall. That is why management suites, which are essentially an integrated set of operations systems, are much more economical in the long term, though they require a very substantive initial investment.
4. Information Technology Infrastructure Library (ITIL) framework
Tools that are built around an IT service management framework, such as ITIL v.3, help an IT organization architect the IT processes and manage IT operations. In other words, ITIL helps design the service around the automation, not just the infrastructure.
5. Knowledge base
A knowledge base essentially transfers knowledge gathered over a long time and from many sources to the target organization. But to be useful, it needs a knowledge-management system as a front end. Knowledge-management systems can be as simple as a store of previous incident notes or a sophisticated database queried by an expert system and integrated with the automation engine directly. In the first case, an engineer will have to search using key words to find the right action to take or the script to run. In the second case, the expert system will use many indicators to navigate decision trees, and either remediate events automatically or present the partial results to an engineer, relieving him or her of doing much of the grunt work required. The expert system also automatically adds to the knowledge base. Unfortunately, at this point, only a few systems targeted toward service providers or large enterprises have attached knowledge bases.
6. Policy enforcement
Many IT organizations develop policies to help IT attain higher degrees of security and quality. But enforcing policies is currently done manually, and is therefore difficult and error-prone. That is especially true when IT staff is under pressure, which ironically is exactly when policies are most needed. Automation engines that have algorithmic enforcement of IT policies offer additional security, compliance and quality improvements.
7. Automation-friendly infrastructure
IT infrastructure with too many systems and interfaces requires more labor to operate and is hard to automate. Worse, infrastructure that requires physical intervention for frequent changes is impossible to automate. So architect your network and storage properly, and virtualize them. If necessary, define them in containers that are optimized for various workloads, tasks or tenants to achieve high levels of uniformity within each container. Some organizations are finding that converged compute systems, with pre-integrated network, storage and compute resources, fronted by one operations system, vastly simplify operations. These converged systems are also very automation-friendly, increase return on investment in the long run, and enable IT to achieve innovation and agility.