How to bridge the IT Ops-CloudOps divide

How do you get traditional IT Ops and CloudOps to work together? Most enterprises struggle with this operations management problem on a daily basis. Generally speaking, the two ops organizations are becoming more siloed—and  more hostile to each other.

Enterprises need best practices that lead to synergies between IT Ops and CloudOps, including shared tasks and the abstraction of monitoring and management above the traditional and cloud-based systems.

For example, you have to monitor security. But how do you do that for both traditional systems and cloud? These should be amalgamated activities, thus the synergy.

Consider these pain points, and then talk about what's working and what's not in your operations organization.

The State of Analytics in IT Operations

Understand the new CloudOps

To understand how to bridge these gaps, it’s important to understand what’s new with CloudOps. Here are a few concepts that differ from traditional IT.

Primary is the notion of working virtually, and at a higher level of abstraction. In the traditional IT Ops world, you always had the option of going to the data center to work on a physical server; this is now impossible with CloudOps. You must rely on dashboards and other tools to understand what's going on, how to fix existing problems, and how to avoid future ones. 

While many people consider this a disadvantage, the opposite is true. Because you work 100% virtually, operations are largely automated and well thought out by cloud computing providers. 

There are two layers with which you work.  

First is the layer operated by your cloud computing provider, where it takes care of the health and well-being of the underlying infrastructure—including hardware, software, and networks—so its cloud stays up and running. This is the primitive layer you never see unless your cloud provider allows you to leverage their physical, "bare metal" servers.  

The second layer is the part of operation that you do see. It's a combination of native cloud dashboards and APIs that you leverage to connect to third-party operations tools. The primitive operational stuff is not out of your hands, which is both good and bad news, depending upon how much operational control you think you need. 

The notion of ops encompasses both on-premises and public cloud. You can think of cloud and on-premises as operational silos these days, but that situation can't persist. 

Problems with the separate-silo approach

Think of users having to call both the cloud help desk and the traditional help desk to get support, and each pointing the finger at the other. If you don’t think that can happen, you're wrong. 

The movement to the cloud won't provide the value you think it should unless security, governance, and operations take the same approach, use the same tool set, and fall under the same organization. CloudOps is often driven by other, siloed organizations with different budgets and leaders. 

"Enterprises are just getting started with cloud services," according to OpsRamp, a cloud services provider. "84% of IT decision makers expect to shift more workloads to the cloud in the near future. While most new apps are being written for the cloud, legacy services are also shifting to cloud." This data means that things will get worse for those who continue to silo operations.

How multi-cloud changes the equation

There's a preconceived notion that multi-cloud makes things worse in terms of operations management. But multi-cloud is actually a good trend for most enterprises moving to the cloud because they can leverage best-of-breed cloud providers, broker services that are the least expensive and that provide the best performance, and hedge their bets in case a provider goes sideways. 

That said, you trade off complexity for flexibility. More often than not, both the IT Ops and CloudOps teams must deal with this complexity. But since they do not yet  play well together, the IT Ops teams deal with them differently than do the CloudOps teams, perhaps in ways that are not yet coordinated between the teams. 

So how does multi-cloud affect traditional IT Ops? Mostly the effect is on the enterprise network and enterprise security. Since cloud providers must all leverage the enterprise network, having more than one cloud can toss more packets on the network than it was built to handle. 

Moreover, you will have data issues, such as enterprise data leveraged by cloud-based applications. Now you must deal with applications that reside on two or more brands of cloud, which complicates security, performance, and even the complexity of doing cost accounting around an on-premises and multi-cloud deployment. 

By running these "concepts" as two different approaches and technology sets, you’re bound to increase your issues in the near term, and systems performance and reliability will likely suffer. 

Again, the solution is for your IT Ops and CloudOps teams to work together, or even become one organization. But that's a reality that most enterprises are not yet ready to face. The transition will take years for most enterprises to achieve, but they need to start the process now. 

Building bridges

Why not just tell CloudOps and IT Ops to work together? This is easier said than done, for a few reasons. Consider:

  • Budgets are not coordinated or combined, including allocations to the cost of cloud versus traditional IT.
  • Each operations organization has its own leadership, which often distrusts the other. 
  • Both ops teams are moving too fast to change now. 
  • The perception is that things are working just fine, but they are not. 
  • There's a lack of cross-training, or the willingness to be cross-trained. 

You need to have a concerted effort in place to drive CloudOps and IT Ops together in a way that will prove successful the first time. 

Best practices

Here are a few core best practices you can follow to drive a high number of successes.

Do a skills-gap assessment

Understand what skills you currently have in both ops teams, consider the skills you need in the future, and understand the gap between the "as is" state and the "to be" state. 

Do this within both teams to identify the current and future skills needs and to understand where overlap exists. You'll find that many on the CloudOps teams have worked with traditional IT operations teams in the past, and many of them may have better traditional skills than those on the IT Ops teams. 

These team members, no matter if on the cloud or the traditional side, become the building blocks for the new "hybrid team" that can handle both cloud and traditional operations management. Most importantly, they must have the capacity to provide consistent ops processes and technology across cloud and non-cloud systems. 

Start organization-transformation planning

Organizational transformation is single hardest thing to do when bridging the ops organizations. But it results in the required end state. 

You need to map out the "as is" and the "to be" in terms of who will work where, what skills are needed, when to train, hire, or fire, and how the operations organizations will be restructured. As a rule of thumb, you should work for a common leadership team and budget for both organizations. Your ultimate goal should be to create a single operations organization for all systems, cloud and non-cloud.

This should take no longer than a year to accomplish. While that goal seems aggressive by Global 2000 standards, you need to move quickly to avoid the cultural issues that tend to spring up around any change. Moreover, large enterprises lose millions of dollars each month by not having IT Ops and CloudOps work together. 

Ongoing improvement

You won't get your bridged or combined operations organizations exactly right the first time around. That's why you need to perform continuous improvement of "combined ops," including continuous improvement of communications, tooling, processes, practices, and approaches. 

Again, this is ongoing, which means you should have a free flow of communication between all of your teams and sub-teams as to new things to try, ways to get better at reactive and proactive operations, and what's changing with the supporting technology.

Keep the endpoint in mind

This approach to converged operations accomplishes a few goals: First, it truly does improve processes and technology for improved operations, and thus improves operations metrics. Second, it gets team members talking who may currently communicate poorly — or not at all. 

I'm not suggesting that you encourage trust falls, but do promote open dialog that is unimpeded by politics or conflicts in goals and objectives. 

It's time to put IT Ops and CloudOps on the same team.