Is Autonomic Computing in Your Future?

By Julie Craig

(Back to article)

Throughout 2006, the high cost of IT support and administration was a focal point for considerable comment and speculation. However, the reasons for this aren't really a mystery. As technology systems become increasingly diverse, distributed and interconnected, they become more complex. Management costs rise with complexity.

IT has become increasingly fundamental to business growth, and as complexity continues to escalate, we, as an industry, have to find better ways to address this roadblock. “Autonomic computing” offers one such solution. It is a concept that builds self-awareness and self-healing capabilities into technology systems, endowing them with the wherewithal to manage themselves and one another.

The evolution of autonomic computing depends on two factors: the willingness of the industry to refocus from human-intensive to technology-intensive support processes, and the ability of vendors to support this shift with innovation. In late 2006, EMA surveyed over 150 IT practitioners at all levels to determine their attitudes towards the first. This article provides some highlights from that research, which was summarized in an EMA paper titled Is the Industry Ready for Autonomic Computing? During late 2007, we will survey vendors to answer the vendor portion of this equation.

Different Names, Same Idea

In 2001, Paul Horn, a senior vice president at IBM, anticipated the technology complexity that has now become a reality. To help mitigate this problem, he proposed autonomic computing. The name comes from the human body's autonomic nervous system, which optimizes such functions as body temperature, heart rate, and digestion based on both internal and external conditions.

Since that time, multiple vendors have launched autonomic computing initiatives. Called "Adaptive Enterprise" by Hewlett Packard and "Dynamic Systems Initiative" (DSI) by Microsoft, autonomic computing is often perceived as an "IBM term" by many vendors. Regardless, the idea remains the same. Vendors are beginning to bring products to market that promote this vision, and these products will likely be the foundation for the "next generation" of enterprise management products.

To evolve into an enterprise management model that addresses evolving business systems, such as SOA's virtualized applications and virtual "devices" that can be provisioned and de-provisioned automatically, increasingly intelligent products are required. We see autonomic computing as filling this gap and as being the hope of the future for cost-effective management of increasingly complex technology ecosystems.

It is fitting that the idea of autonomic computing came out of IBM, a bastion of mainframe and midrange computing, because autonomic- or self-managing- capabilities have been built into hardware for many years. However, just as the IT Infrastructure Library (ITIL) is leading the way for IT support organizations to become less siloed; that is, less divided into specific technology verticals, autonomic capabilities are expanding beyond the hardware realm to encompass not simply devices, but entire technology ecosystems. The long-term vision is of interconnected, interacting IT systems that, via standards and cross-communication, are able to self-regulate and adapt to their environments.

Executive Attitudes

In late 2006, EMA surveyed personnel at all levels of IT organizations, from CIOs to technology administrators. The research was designed to gather user attitudes and comfort levels towards various autonomic capabilities.Although respondents came from companies of all sizes, the largest single segment of respondents were employed by enterprises. Twenty four percent worked for companies with revenues of $10 billion or more per year. Service providers, financial organizations, government and the insurance industry accounted for 42% of the respondents. Approximately 62% were C-level executives, architects, directors or managers, with the remaining 38% representing hands-on technical personnel.

The survey yielded some interesting results. When asked to rank the top three concerns affecting their departments from a list of ten, both groups agreed on three of the same concerns. The high cost of supporting IT, lack of alignment with the business, and lack of management products were top concerns for both groups. Executives and technical staff did differ, however, in two significant areas: Thirty one percent of the executives cited lack of technical expertise as their number three concern, versus none of the technicians. This likely reflects the executive perspective to end-to-end business services, versus a more siloed-technical view. Thirty nine percent of technical staff cited project backlogs as tying for the number two position, versus none of the executives.

Seventy two percent of executives verses 58% of technical personnel, however, were familiar with autonomic computing concepts. When asked whether their organizations use ten specific autonomic features, the two groups differed considerably. Products that detect and mitigate security threats (19%) were the "most used" autonomic feature cited by executives, while 31% of technologists cited solutions that "automatically trigger actions in IT environments and notify support staff of the actions taken.”

Willingness to use autonomic technologies was, in general, higher among executives than among technologists. For example, more executives than technicians would be willing to deploy threshold-based provisioning and de-provisioning, products that re-prioritize batch jobs based on workload, and products that automatically launch fixes for "known problems" when they are detected. On the other hand, technologists were more willing to use products that calculate and track rolling baselines and products that trigger actions in IT environments.

It was also clear that both groups equated autonomic technology with improving the levels of service being delivered. Of eight possible drivers for using autonomic technology, the two clear winners were "improved customer satisfaction" and "improved alignment with business goals". This response far outstripped the runner-up, "improved performance against SLAs".

This research produced some very interesting results. It demonstrated that IT technologists, typically viewed as heads-down "geeks" with little perspective to the business, are as concerned with IT costs, customer satisfaction, and business alignment as their management. It also indicated that they are much more accepting of autonomic technologies than anticipated.

The key for generalized acceptance appears to be audit trails. IT wants autonomic products to create audit trails of the actions they take, to provide a human readable history of automated actions. This is in keeping with the general shift to control and governance that has been prompted by best-practice methodologies and regulatory requirements.

Executives, as expected, looked to the cost benefits of autonomic computing, including the ability to support increasingly large technology deployments. A big surprise was that both groups cited "lack of tools" as a concern, especially since separate EMA research found that the average company has between five and 25 enterprise management products in place. This indicates a gap between the functionality available within their current management products and the functionality they need to support the business.

Julie Craig is a senior analyst with Boulder, Colo.-based Enterprise Management Associates, an industry research firm focused on IT management. Julie can reached at jcraig@enterprisemanagemengt.com.