Originally published on The Enterprisers Project.
Whether you go hybrid cloud “accidentally” – as many enterprises do – or intentionally, you need a strategy to increase the odds of its success and measure results. Apply this expert advice on how to build a strong hybrid cloud strategy.
Many organizations become hybrid cloud shops unintentionally. You add this environment for one particular initiative project, or integrate those two systems because of a merger, or any number of other possibilities.
Next thing you know, you’re operating a hybrid cloud by most definitions of the term.
“Many organizations fall into hybrid cloud through some combination of acquisitions, lack of coordination among lines of business, or just a lot of ad hoc actions taken on a project-by-project basis,” says Gordon Haff, technology evangelist at Red Hat.
Hybrid cloud strategy: 5 keys to success
This leads us to a raison d’etre for this article: Whether you go hybrid cloud in an “accidental” manner or by intentional design, you need a strategy to measure and boost the odds of its success. In the unintentional category, that strategy may have to be put in place after the fact, and that’s fine. What’s important is that you do have an adaptable plan. And that’s first on the list of must-haves.
1. The most essential part: Develop a strategy
“Perhaps the most essential element of a good hybrid cloud strategy is actually having a strategy in the first place,” Haff says.
The accidental path to hybrid cloud is relatively common, especially as IT leaders and teams perform a challenging balancing act between fulfilling today’s business needs while also building for the future. That means that it’s never a given that an organization has actually thought through their long-term strategy. Again, it’s OK to retrofit a strategy to assets and architecture that have been put into place in a more one-off or project-based fashion (and it’s certainly better than ignoring strategy).
“While unexpected events and needs are a given, one of the roles of IT is to rationalize the company’s portfolio of cloud services and data center resources, which includes making course corrections as needed especially as overlapping or duplicate capabilities accrue over time,” Haff says.
2. Strike the balance between old and new
Here’s another balancing act your hybrid cloud strategy will need to consider: How will you make decisions about what to run where? By definition, a hybrid cloud environment will likely have a mix of both cloud technologies and what some might call “legacy” technologies, such as bare metal servers on-premises, or monolithic applications that run in a colocation facility. Perhaps a better word than “legacy” (or “old”) here is “existing” – as in, existing assets and investments. Part of the hybrid approach’s appeal is that you don’t have to dump those to benefit from newer cloud platforms and services.
Moreover, not all organizations are necessarily interested in keeping up with the rapid pace of change in cloud and cloud-native platforms.
“Businesses planning on implementing a hybrid cloud environment need to consider their appetite for managing the constant evolution of the cloud landscapes,” says Ryan Murphy, VP and North American cloud center of excellence leader at Capgemini. “Technology innovation and obsolescence are the primary driving factors for change.”
Another hybrid cloud benefit is that a particular organization retains more control over the pace of that change. You don’t always have to be running on the latest release of cloud service X just because a vendor wants you to be, nor are you stuck entirely with aging infrastructure. You can achieve a balance across environments, but there will be some necessary decision-making about what runs where, how often you’ll move workloads between different environments, and so forth.
If you’re doing it right, you’ll make these decisions based on the criteria that matter most to your business, whether these include cost, performance, security, compliance, resources, or others. “All of the above” is an acceptable answer too, as long as you’re able to prioritize criteria that might come into conflict at different points.
Containers and Kubernetes can also give you added control over workload portability, as we explored in our related article, 3 reasons to use an enterprise Kubernetes platform. We’ll get back to this topic in tip #5.
You’ll need to define how you’ll know when newer is in fact better. A straightforward example from Murphy: If “protecting existing investments” is code for something like “running on a server with an unpatched or unsupported OS installed,” then you’ve probably taken shortcuts on understanding the true costs (and in this case, the security and compliance risks) of your what-to-run-where framework. Your strategy should address these kinds of concerns and the criteria you’ll use to make “existing versus new” changes.
3. Address chances to improve process and culture
“Existing versus new” is not just a technical consideration.
Daniel Herndon, senior product manager for cloud infrastructure at Laserfiche, notes that many hybrid cloud strategies rise out of larger transformations. Regardless of the catalyst, however, your strategy should examine other areas for improvement or updating, and not just in the technology portfolio: Think process, culture, and so forth.
“Any good hybrid cloud strategy should be preceded by an extensive audit of existing processes, solutions, and systems,” Herndon says. “A hybrid system will provide an organization the new opportunities that come with new technologies, solutions, vendors, and industry best practices.”
Use your hybrid cloud roadmap as a chance to identify and increment wins elsewhere; perhaps beginning to build a CI/CD pipeline in phases and in the process, shifting the organizational mindset around testing, security, and other areas that might have historically been treated as functional afterthoughts.
The general idea is for your strategy to be connected to the bigger picture of your IT team and the larger organization in which it operates.
“A good hybrid solution isn’t just about changing where workloads are [run], but also should support the organization’s innovation in thinking, practices, software, and technologies,” Herndon says. “You wouldn’t put old tires on your new car to save a little money.”
4. Plan to measure and manage major variables
Another key area for your hybrid cloud strategy is to define how you will weigh the benefits – real or perceived – of your strategy in a business-appropriate context. Meaning, it’s not just enough to say, “We can scale up and down faster now because we moved workload X to cloud Y now.” You also need to be able to measure that against significant variables such as costs and performance, and make sure your use cases make sense. A predictable, relatively static application, for example, might be better served on-premises.
“Cost and performance are not constants,” says Scott Sneddon, senior director and multicloud evangelist at Juniper Networks. “There are many factors that cause variability in these areas.”
Costs can go up and down in the aforementioned scaling up (and down) scenario, for example. Sneddon also notes that costs can also be driven up and down with network traffic growth and as storage needs change.
Performance and performance requirements, meanwhile, change depending on the load of a given application, the time and location of a given application, and the load on the underlying infrastructure supporting that application, Sneddon says. “The need to adjust an application’s location and deployment based on these variables must be taken into consideration when choosing cloud platforms and the tools used to operate those clouds.”
And you can shift your vantage point here: As mentioned above, “protecting existing investments” should be exactly that, not a cover story for sacrificing areas like performance or security solely in the pursuit of cost reductions.
“An IT organization needs to understand the true costs to the business when operating solutions on slower infrastructure, or when utilizing operating systems that are either no longer supported, or don’t take advantage of the newer technology innovations,” Murphy says. “Businesses need to understand the agility available from each element of their hybrid cloud operating model in order to ensure their workloads are located where they make sense from a security and financial perspective.”
5. Identify open tools that work across your environments
Another critical facet of your strategy: What tools will you use to build and manage everything? What do you already have in place, and what will you need to add? Where are the gaps?
In particular, Sneddon says, using standardized, open tooling that will work across multiple platforms or environments is crucial. This is a good example where a less intentional shift to hybrid cloud can cause problems: If you don’t, you might find your future decisions are hampered by a lack of flexibility.
“Enterprises need to avoid lock-in by using standardized tools wherever possible,” Sneddon says. “And choosing simplified and standardized network and security tools – that can also be used in the private cloud and across the rest of the enterprise – can significantly reduce the risk of lock-in to a single cloud platform.”
Here’s an example that gets back to those major variables: As performance needs for an application change over time, there are some moves you can make within the same cloud platform, such as changing the machine size or changing the location where that workload is running within the same cloud ecosystem. But what if you determine that workload’s performance would be better served in a private cloud? You could be in for some pain if you haven’t considered this scenario in advance as part of your strategy.
Sneddon shares some parting tips on this front in terms of addressing tooling as part of your hybrid cloud strategy:
- Choose a standard and open cloud platform like Kubernetes.
- Choose simplified and open networking and security platforms that can provide commonality and consistency across private and public cloud.
- Choose a monitoring and analytics tool that can instrument your entire cloud footprint, as well as the network and security components. This will enable an end-to-end view that can greatly reduce troubleshooting time.