kran77 - Fotolia

Autonomic computing may pave the way to software-defined data centers

As IT shops strive for competitive advantage with their web-based applications and services, an increasing number are turning to various flavors of autonomic computing.

IT shops striving to deliver more competitive cloud-based products and services are dawning on a new realization: Delivering these offerings is only the start.

As more such projects get underway, IT pros are beginning to understand that they also need to reshape the underlying IT infrastructure to optimize these services and improve the management of existing IT assets. The answer to these new problems might be buried in the past: autonomic computing.

The growing number of enterprises adding multiple clouds for different services requires a tangle of on-premises and cloud-based infrastructure. More than 70% of organizations currently have multicloud strategies in place, and that number will rise to 90% over the next couple of years, according to IDC.

"There is this big shift in IT now where they see themselves more as service providers, where they have to mix and match services from multiple clouds to meet various workload requirements," said Mary Johnston Turner, research vice president at IDC's enterprise systems management software group. 

While both small and large vendors scramble to piece together products and services to address these shortcomings, some IT organizations still crave something more comprehensive from a single source to simplify everything. That option, unfortunately, doesn't seem likely anytime soon.

"There's not a single tool out there that helps us with application performance, manages assets across clouds or automatically configures things like VMs on the fly," said a software engineer with an insurance company in Hartford, Conn. "It's hard to piecemeal a solution together that can handle that many moving parts."

There's not a single tool out there that helps us with application performance, manages assets across clouds or automatically configures things like VMs on the fly ... It's hard to piecemeal a solution together.
Software engineer, insurance company

As IT pros struggle to piece together an enterprise-class tool chain that can handle many moving parts, analysts are sympathetic.

"Few vendors have the IP needed to address this rate of change," IDC's Turner said. "The lines have blurred between development and infrastructure and between public cloud subscription services and on-premise[s] operations. [IT shops] need solutions with an awareness of application performance and that understand the dependencies of the whole stack from the code to the infrastructure."

This lack of a complete package explains why so many software vendors have snapped up smaller vendors or formed partnerships with others to acquire competency where they have none.

To cite one example, BMC and Turbonomic cut a deal in October 2016 where BMC will resell Turbonomic's autonomic cloud platform, which is designed to help heterogeneous environments self-manage.

Products such as those by Turbonomic and others automate lower-level yet time-consuming decisions of systems administrators such as resizing virtual machines, when and where to deploy a container and orchestrating a variety of different applications and infrastructure -- ultimately to more efficiently use IT resources and assets, thereby eliminating the need to buy new hardware or software.

An old term for a new data center

These companies want to reintroduce a concept from the 1990s called autonomic computing, which describes a self-managing, self-healing computing model that controls the functions of applications and hardware systems without any intervention from IT administrators. The idea is that autonomic computing will bring customers closer to software-defined data centers (SDDC).

Given the deep saturation of virtualization in most data centers, plus the trend toward SDDC, some say the stage is set for autonomic computing to make a comeback, albeit in a different context.

"The idea of an autonomic data center has always made sense to us, but the underlying and fundamental technology enablement for it was not in place until recently," said Charles Crouchman, Turbonomic's CTO.

The key to advance the state of the art is with control systems, according to Crouchman. "We are [accustomed] to batch analytics, and using that with autonomic computing in today's data center would be akin to your nervous system working once every 24 hours. To function, your nervous systems needs to work continuously," he said.

Another indicator of the growing interest in this market was an October 2016 deal between Zenoss Inc., a provider of IT monitoring and analytics software, and SaltStack, which makes orchestration software for systems management. The combination of respective products aims to better enable IT shops to create an autonomic software-defined data center.

The combined products will be able to scale data center capacity based on system events, and trigger intelligent system remediation by event creation and consumption.

"The CIOs we talk to can see that their businesses are relying more and more on health and performance of their applications and services," said Greg Stock, CEO at Zenoss. "So they are demanding more and more that the technologies they choose can keep pace with that."

A somewhat different approach to automating management of IT resources is ServiceNow's ServiceWatch Suite. The offering works with the company's configuration management database, which consolidates all monitoring data to a dashboard to present all business services in a single view. The products also manage public and private clouds, and include self-service provisioning and event management.

When IT shops need to modernize their operations to be more competitive, "they can't move fast enough without that automation in place," said Farrell Hough, general manager at ServiceNow's IT service management business unit.

One example where the lack of automation hurts enterprises is in the difficult task of moving local applications and data, which typically carry a high total cost of ownership and are difficult to manage, to the cloud.

"Almost 99% of the user base is taking some advantage of the cloud and would be happy to relinquish control of the [on-premises] infrastructure, which they spend so much time managing," Hough said. "But automation can lessen the burden of that lower-level, repeated work so they can focus on innovation."

IT shops reluctant to invest in automated software do so because they either don't want to spend the money, or they don't have the skills in-house. Some, however, say they are actively exploring the ROI benefits of the technology.

"I can see where technologies like this [autonomic computing] will find a place in many larger companies," said one IT professional with a large manufacturing company in Minneapolis. "Technologies like this touch so many parts of a data center, we need time to explore what sort of disruption it causes and what the costs are for adding to or replacing existing on-premise[s] and cloud products that work with it."

Ed Scannell is a senior executive editor with TechTarget. Contact him at [email protected].

Next Steps

Discover open source tools to handle diverse systems management

How site reliability engineering dismissed administration

IT administrator skills in an autonomic age

Dig Deeper on Application Maintenance on Production Systems