You are here

4 things you need to know about managing complex hybrid clouds

public://pictures/davidl.jpg
David Linthicum, Chief Cloud Strategy Officer, Deloitte Consulting

Setting up a hybrid cloud involves more than just pairing private and public clouds. Hybrid clouds are usually complex, hard to build, and hard to manage. The good news is that best practices are emerging; you can now learn from the mistakes of others. The bad news is that complexity continues to grow, and managing hybrid clouds will only get more challenging. 

If you already manage a hybrid cloud environment, or soon will be, here are four critical guidelines for deploying and managing complex hybrid clouds.

1. Simplify your domain 

Think in terms of small cloud-services clusters that can be managed at the cluster level or, holistically, using tiers. The idea is to break up your architecture into manageable domains. For instance, if your hybrid cloud spans three organizations in your company—say HR, production, and sales—then perhaps it’s time to break up that hybrid cloud by organization. 

While you may share the same hardware and software, as well as a public cloud or two, the concept is to simplify management by segmenting a complex hybrid cloud by application groups, such as those that support HR, production, and sales. This results in separate domains that are easier to manage as simplistic segments with fewer workloads. 

Many people who manage clouds, including hybrid clouds, think this makes things more expensive because you need separate management teams and toolsets for each domain. That's not always the case. Yes, you need to manage each domain separately, but you don’t need separate teams to focus on each domain, and you can leverage management tools that span domains. 

The guideline boils down to: "When in doubt, break it out." Moreover, break it out using logical separators, not physical; use workload types, organizations, or other common patterns to find the separate domains. 

[ Digital transformation can be a costly failure without proper controls. Find out how IT4IT value streams can help in this Webinar. ]

2. Don't mix security and management

Most people who manage hybrid clouds understand the differences between security and management, but too often those two are mixed. The reasons are usually cost-related, but the typical outcomes are higher cloud security risks and non-effective cloud management. 

Security needs its own set of owners. It's a matter of two different approaches and toolsets that can be in conflict with each other. 

For example, hybrid cloud security systems affect performance, and hybrid cloud management typically attempts to enhance performance. The result could be that the shutting down of security subsystems in order to provide better workload performance results in a breach months later due to the vulnerabilities the measure introduced. 

Hybrid cloud security and hybrid cloud management teams should be closely coordinated. But they also should be separate and peered organizations, meaning that one should not report to the other. That's a common mistake. 

Also, security and management should not be at odds with each other; each has separate and distinct objectives. Management optimizes the hybrid cloud in terms of performance and reliability, while security reduces the risk of a breach. Each has its own set of tools and approaches and keeps separate talent pools, so you can maximize both management and security at the same time.  

[ Looking to bring innovation into your enterprise? Learn from others' Enterprise Service Management (ESM) implementations—and get recommendations for deployment. ]

3. Allow only qualified people to manage your hybrid cloud

The stories of failure here are numerous. Now that there's an industry-wide cloud computing talent shortage, IT organizations are compromising. That means sometimes underqualified people get important jobs, such as hybrid cloud management. You can easily predict the outcomes from such hiring practices:

  • Breakdown of core hybrid cloud services, typically around performance and reliability issues. Failure to automate the management tasks can result from using the wrong toolsets, or no toolsets, so the hybrid cloud will suffer along with its users. 
  • Lack of hybrid cloud management planning, including gathering the right budget for people and technology. Those who manage a hybrid cloud need to plan at least two years out, and should understand the growth that the hybrid cloud will experience, as well as the business and technical requirements. 
  • An inability to respond to a crisis. The public cloud goes down, the data center experiences a power failure; these are things that bring down people who lack the required talent, training, or experience. Most issues that have occurred over the last three years related to hybrid clouds can be traced to a lack of the talent required to properly respond to the issue. This is why minor issues quickly expanded to major ones, and the damage caused far outpaced the amount saved by cheaper hiring.   

On the other hand, if IT leaders can’t find qualified people, what are they to do about hybrid cloud management? It's really a matter of what you're willing to pay. And while cost is unnaturally high for many of the best candidates, compromising could cost far more in terms of efficient delivery of services and the ability to respond to a crisis. 

4. Get dev and ops together, already 

DevOps is an essential part of making a hybrid cloud work, but not all enterprises have bought into it. Some continue to keep development and operations in separate domains that do a poor job of communicating and fight often. 

The benefits of DevOps are obvious and well documented, and relate directly to hybrid cloud management. These include:

  • The ability to continuously improve hybrid cloud management approaches and technology by interacting with the developers and automated tools that they use. This advantage cannot be overstated. Those charged with hybrid cloud management capabilities can gain direct knowledge about what's working and what's not, in terms of how developers approach the development, configuration, testing, and deployment of workloads. This feedback cycle makes developers and hybrid cloud managers better at understanding what needs to change to improve management.
  • The ability to build hybrid cloud management services directly into the applications. While many who manage hybrid clouds believe their jobs stop at the application workload, that’s not the case. Indeed, "hooks" and APIs can be placed in these workloads that allow those charged with hybrid cloud management to manage the workloads as just another resource. So, while you can manage storage and other resources within your hybrid cloud, you can manage application workloads as well, plus the workloads that interact directly with automated hybrid cloud management tools. 

DevOps is perhaps the most important concept mentioned here. Consider the impact it has on hybrid cloud management, and on the quality of your hybrid cloud service overall. Those who fail to see the links, or even attempt to separate dev and ops, often find that hybrid cloud management suffers the most—including the hybrid cloud users, and thus the business. 

Don't blow it now

So, where are enterprises in the adoption of hybrid clouds and hybrid cloud management? They are just getting to the end of their cloud migration projects, and have now moved into the management and operations stages. This means that they need to get good at hybrid cloud management, and fast. All the planning, migrations, and deployment successes won’t matter if a poor-performing hybrid cloud, or one that has reliability issues, is the result.  

Of course, nobody wants to fail, but few best practices and technologies exist, and they are not well-understood by enterprises that are the proud new owners of a hybrid cloud. A bit of trial and error is going on right now,

Your core strategy should be to follow the best of the best practices. Plan early, and change the plan as needed. A comprehensive plan includes:

  • Following steps to set up hybrid cloud management.
  • Allocating resources and the budget required.
  • Creating job descriptions and sourcing plans.
  • Selecting and implementing the required technology, as well as test and acceptance plans.
  • Creating implementation schedules, as well as a business case.

That said, this is not a guarantee of success. You're merely ensuring that you’ll have the greatest chance of being successful. This is about following the best of the best practices, and avoiding common mistakes. Not everyone doing hybrid cloud management understood this, but why not learn from their mistakes? 

[ Ready to manage your hybrid IT future? Download Crafting and Managing Hybrid Multicloud IT Architecture to get up to speed on unified infrastructure management. ]