F5 Blog

 
Archive Search Articles

Rigidity of the Org, not Hardware, is Holding up App Deployments

The solidification of the divide between traditional IT and a more agile, DevOps-driven IT is resulting in the emergence of essentially a two-tiered, data center architecture. This should not be surprising to those paying attention. For years I’ve been inundated with questions from operations and developers who’ve wanted to move faster, more frequently, but were stymied by the rigidity of not hardware in the core network, but of the processes that require manual approvals and processes to implement changes. the network still in the way

The friction that naturally exists between those tasked with maintaining secure, stable, and fast transfer of data in and out of the data center and those tasked with getting to market – now – those applications and features believed to be the catalyst for business growth causes a lot of heated words, finger pointing, and bad feelings.

Many like to blame “boxen”. Those aging, legacy pieces of hardware that reside in the network and, in spite of providing the critical services applications need to scale and stay secure, are just “getting in the way.” But the reality is that those boxen are not to blame themselves. Almost all today sport an API that enables anyone with access to curl and a working knowledge of HTTP to configure, manage, and otherwise manipulate the services that accompany every application deployment. Many of these same “boxen” are virtualized themselves, and serve to provide the same kind of virtual experience as that deeper in the data center.

They are hardly roadblocks on the expressway that is the notion of continuous deployment.

Rather, it is the rigidity of organizations, in the culture, as DevOps proponents are wont to say, that causes this friction and slows down application deployments. Because there is a need to maintain stability and security in production that does not necessarily exist in the more volatile environment that is app dev, there is less tolerance for mistakes and the inevitable disruption to business they cause.

This is not necessarily a bad thing. The core network certainly needs to maintain a high level of availability, not just to ensure access to applications inside the data center, but for those users inside who rely on public, SaaS solutions to do their jobs.

Organizations need to relax constraints, however, in order to better balance the needs of both architectural tiers in the modern data center. App deployments do need to happen faster and more frequently. To achieve that organizations need to take a serious look at the processes currently in place that are almost certainly more rigid and unyielding than the boxen they govern. IT operational processes are as important today to business as its business processes, the ones that business leaders understand must be examined, evaluated, documented, automated, and improved to yield efficiency and business performance benefits. Even more importantly, they need to be agile enough to be changed when necessary. But to do that, organizations need to understand those processes and be willing to evaluate them with a critical eye toward inefficiency and chokepoints.

In a 2016 AIIM Industry Watch Report titled “Information Management – State of the Industry 2016”, fifty-five percent of respondents said Business Process Management (BPM) was significant (38%) or imperative (17%) for their business.

When it came to the use of technology to augment processes, businesses saw improvements across the board:

❶ One-third of organizations have seen a decrease in their review and approval cycles and 62% say they have gained benefits from better routing to and between individuals. For 42%, the benefits come from greater organizational agility and routing between processes.

❷ Faster processing of business critical activities has provided the biggest value for 53% of respondents. 45% cite their biggest value from fewer errors and exceptions processing.

❸ Payback has been realized within a year for 41% of respondents of which 17% cite ROI within 6 months. For 25% their ROI was within 18 months.

IT uses the term “orchestration” often but rarely discusses how it relates to process. It is when orchestration is used to automate operational processes that govern deployments that real improvements will be seen. It is discovering waste and gaps in the value stream and eliminating wait times and unnecessary approvals and redundancies that will net organizations the same kind of gains in velocity and agility that they have seen on the business process side of things.

We can call it DevOps or we can call it Operational Process Management (OPM), but the reality is that the two are closely related. Organizations need to do to operational IT processes what they’ve already done with business processes to achieve greater agility and speed of app deployments without sacrificing stability and security.

And has to very little to do with whether there are boxen or not in the data center.