How the cloud fractures application delivery infrastructure ops

The problem with fractured operations

The traditional infrastructure team still operates ADCs and load balancers in the data center, while preferring the vendors they have worked with in the past. DevOps and CloudOps have taken control in the public cloud, choosing to use software and cloud provider services that are more integrated with their DevOps toolchains.

This fractured operations model is problematic. Companies with divided Layer 4-7 operations are less likely to be successful with this infrastructure. EMA research participants also revealed why they feel a need to close this operational gap.

First, 43% of enterprises said this situation has introduced security risks. In most enterprises, application delivery infrastructure is an important component of overall security architecture. Companies need to take a unified approach to network security.

Research participants identified compliance problems (36%) and operational efficiency (36%) as the top secondary challenges associated with fractured operations. And 30% said platform problems -- such as issues with scale, performance, functionality or stability -- are a major challenge.

Closing the operational gap

EMA doesn't have a guaranteed fix to this issue, but we do have some suggestions. First, enterprises need to recognize the problem. IT leadership needs to tell groups on either side of this operational divide that they should work together to unify their overall approach to application delivery infrastructure. This doesn't mean one team's tools and platforms should be the de facto services for everyone. It simply means these teams should acknowledge the need for a consistent approach to application delivery management.

Next, the IT group should select standard application delivery infrastructure platforms, along with standard configurations, policies, and operational tools and processes. These standards don't have to be one size fits all. An organization could identify two or three standard platforms -- depending on the requirements of individual teams -- and build policies, golden configurations and toolchains around each of these platforms.

The key is standardization, not prohibition of the best platform for a job. EMA's research found that 81% of enterprises that are trying to close the operations gap have standardized on one or more application delivery platforms. Another 16% would like to standardize.

Finally, vendors need to step up. They have to recognize they can't expect to win the loyalty of all the different application delivery infrastructure buyers inside an organization.

EMA asked research participants how vendors could help them unify operations. The most common responses were deployment flexibility (53%) and improved APIs for integration (53%). Successful enterprises were more likely to single out APIs as a critical need. Nearly half (49%) wanted their vendors to provide multivendor support on their management tools, and 48% asked for more flexible licensing.

It's never too late, nor too early

Regardless of where an organization is on its cloud journey, IT groups should be thinking about this issue. Whether one is in the early days of migrating applications to the cloud or the last days of retiring webscale ADCs in the data center, network teams and DevOps teams need to communicate and collaborate.

It is fine if each team has its own approach to application delivery infrastructure, but those separate paths should have some coordination so they don't conflict with each other and don't present risks to the business.

Don't let silos rise up between your management domains. Standardize on as small a number of platforms as possible, and establish best practices and policies around these technology stacks. Also, keep lines of communication open between these groups, and adopt operational tools that provide end-to-end visibility and control across different application delivery platforms.