4 signs that your hybrid cloud strategy is wrong (you do have one, right?)

Edge-to-cloud, built to transform your business. Learn more about it here.

There is endless talk about hybrid cloud and how it is the way forward for many organizations. There’s a growing consensus that neither pure public cloud nor a wholly on-premise setup is the best way to run an IT infrastructure: it’s time to mix the two.

The first mistake organizations make is thinking that the decision to go hybrid is all that there is to do. There are numerous things that can go wrong with a hybrid implementation, many based on false assumptions. Here, we examine four signs that your hybrid cloud strategy is wrong – and reveal what can be done to ease the pressures.

Management is too complex

One immediate issue facing cloud management is that workloads could be running in different places. In particular, there’s a constant challenge when trying to balance compute instances in an on-premise datacenter with instances deployed in the cloud.

Matt Johnson, CEO of cloud specialist Rayo, told us that management is the hardest thing for cloud providers to do. “The cloud providers are offering more and more but still don’t have the control of on-prem,” he said. “They will offer ‘single pane of glass management’ but are dependent on a single cloud provider.”

One way to get to grips with hybrid management is a comprehensive tool. HPE Morpheus Enterprise Software is such a product, offering the ability to enable on-prem private clouds, centralize public cloud access, and orchestrate change.

There’s a cost issue too, added Johnson. “A common mistake in implementing hybrid cloud strategies is underestimating how unpredictable costs can become when moving from on-premises infrastructure (with mostly fixed costs) to variable cloud pricing models.”

You’re struggling to work out where legacy IT fits in

Most companies will have an array of legacy kit to work with. In extreme cases this may involve extremely old mainframes that have been running for decades, but usually it’s a range of hardware on three-year (or longer) refresh cycles.

Any business implementing hybrid cloud must think about how to work with this kit and tough decisions will have to be made.

Even cloud customers will have their own legacy issues to worry about. “You can get legacy in the cloud – there may be different cloud offerings to bring in,” said Johnson. “There are as many as eight generations in some cloud products.”

Stewart Parkin, CTO of Assured Data Protection, agrees that management complexity and the need to incorporate legacy kit will always present a challenge. “Hybrid cloud promises flexibility and agility, but too often organizations underestimate the operational complexity that comes with it,” he said. “Managing multiple environments, integrating legacy infrastructure and trying to rein in unpredictable virtualization costs can quickly derail the strategy.”

Parkin added that companies going down the hybrid route need to develop a completely different mindset when it comes to incorporating legacy kit. “A common mistake is treating the hybrid cloud as a simple extension of existing IT, when in reality it requires new thinking around workload placement, governance and cost transparency.”

It is possible, however, to deploy a set of tools that will enable systems managers to migrate their legacy kit efficiently. HPE Application Modernization Services allow organizations to transform their systems effectively, ensuring that legacy hardware is integrated fully into a hybrid system, continuing to function efficiently but supporting the new hybrid cloud services.

Infrastructure Meets AI‑Driven Scale

When AI moves to production, the infrastructure becomes the bottleneck. The piece outlines how organisations can optimise for throughput, lifecycle management, observability and cost‑efficiency, by aligning their infrastructure stack with the needs of models in continuous operation, not just experiment mode.

If you already have your models ready, this article helps you map your stack for real‑world readiness

Your virtualization costs are unpredictable  

One of the initial attractions to cloud computing was that it would lead to great cost savings as organizations would only pay for what they used. This quickly proved to be a misconception for a variety of reasons.

For starters, the sheer ease of use meant that cloud instances could be started but often left running. Then there was the unpredictability, and complexity, of cloud pricing. To a cynical IT director, it seemed like every cloud provider had a variety of costs and used different pricing models.

This was one reason why many organizations switched to private cloud models. Here, they could buy servers and know the fixed cost. The issue was predicting capacity. Every time an organization needed to increase its capacity, whether compute or storage, there was the need to provision new hardware. This cost money, of course, but also time.

A hybrid system will solve some of those issues but there will still be uncertainty over costs.

Johnson said that this is an issue that should be factored in from the outset. “The key is to treat cost as an architectural consideration (this approach is sometimes known as cost-conscious design), with your choices directly influencing the bottom line,” he said. “You should set sensible scaling limits and throttling thresholds, and implement detailed cost monitoring.”

Once again, HPE is also well positioned to help. HPE Private Cloud and HPE Private Cloud Business Edition take the pain out of provisioning cloud services and also simplify virtual machine management, whether the VMs are run on-premise or in public clouds.

You don’t have control over where workloads are placed

Cloud customers used to have little idea where their workloads were being placed, which caused several problems. Organizations faced issues over latency, security, regulation, and sometimes a mix of all three. However, with most modern public cloud providers, customers do have some understanding of where their workloads are.

This does mean that organizations must be aware of the policies of the provider. They must also draw up watertight contracts to ensure that they have some level of control, particularly regarding security and sovereignty. If in doubt, there is always the option to keep anything particularly sensitive on-premise.

Conclusion

Going down the hybrid route requires clear thinking. It’s not a question of firing up a bunch of cloud instances and running them through your existing infrastructure: you must map everything out.

“The organizations that succeed are the ones that build a clear management framework from day one, put guardrails around legacy systems, and adopt tools that provide visibility and predictability of spend,” said Parkin. “That way, hybrid cloud becomes an enabler of resilience and innovation, rather than another layer of IT headache.”

Fortunately, companies such as HPE have a range of solutions in place to make management of hybrid infrastructures easier to handle. But all IT managers should be aware that there’s no substitute for having a clear transformation plan in place.