This article draws on industry analysis originally published by Gigabit Magazine (gigabitmagazine.com), featuring insights from Eran Brown, CTO of Infinidat EMEA.
What Is Hybrid Cloud and Why Does It Matter?
Cloud computing sits at the heart of most digital transformation programmes — alongside IoT and AI/analytics — yet it is often the least visible in day-to-day operations. Without an integrated data infrastructure, the scale of connectivity and optimisation that transformation promises is simply unachievable. Hybrid cloud pushes the technology further by letting organisations synthesise multiple cloud models — both private and public — into a single, purposefully designed solution that addresses their specific workload, compliance, and cost requirements.
Public Cloud Is Not Always Enough
The wide availability of public cloud platforms — including Microsoft Azure, Google Cloud Platform, and AWS — means most organisations can secure a package that covers basic storage and compute needs. Each provider offers hundreds of features, and some have built industry-specific tooling on top of their core platform (IBM Cloud’s Watson for IoT is a frequently cited example).
However, a single public provider is not always sufficient. Data volumes, regulatory constraints, latency requirements, or the need for dedicated infrastructure can push organisations toward building or leasing private cloud capacity. Once a private cloud is in place, connecting it to one or more public clouds over a sufficiently capable WAN creates the hybrid architecture.
How to Evaluate a Hybrid Cloud Strategy
Eran Brown, CTO of Infinidat EMEA, argues that the evaluation of any hybrid cloud model should focus on four practical dimensions: agility, flexibility, speed-to-market, and cost efficiency. In his words:
“Companies will need to draw a strategic line between how and when they use public and private cloud for data storage. It will be impacted by the quantity of ‘legacy’ data to be stored, as well as what data companies want to ‘own’ or store locally; not all data is suitable for storage in public clouds.”
Eran Brown, CTO, Infinidat EMEA
This means that before committing to a hybrid model, teams should complete three preparatory steps:
- Define the desired outcomes. Be explicit about what you expect from the hybrid arrangement — lower latency, data sovereignty, cost reduction, or burst capacity.
- Map data placement rules. Identify which data sets belong on private infrastructure (sensitive, regulated, or high-churn data) and which can safely reside in a public cloud.
- Model the financial impact. Understand both short-term spend and long-term total cost of ownership before signing contracts.
“Choosing between on-premises and public clouds should not be driven by hype, but rather what will enable units at a cost that is acceptable for the long-term viability of the business.”
Eran Brown, CTO, Infinidat EMEA
The Real Economics of Cloud Storage
Public cloud storage costs are rarely as simple as a headline price per gigabyte. A survey by storage vendor Wasabi (cited in the original Gigabit Magazine article) estimated that customers pay approximately $0.46 per GB/month on Azure, $0.26 on Google Cloud, and $0.23 on AWS — though these figures exclude regional pricing variances, egress fees, and API request costs, all of which can significantly increase real-world bills.
For workloads involving high-resolution video, large image libraries, long-form documents, or extensive transactional archives, these per-GB charges accumulate quickly. As Brown notes:
“As data growth continues unabated, larger organisations will be under increased pressure to more clearly understand the ‘cost vs benefit’ for different storage models as the volume of data expands exponentially.”
Eran Brown, CTO, Infinidat EMEA
On-premises private cloud hardware can offset these costs for data-intensive workloads. Storage vendor Cloudian has published a report claiming that on-premises storage hardware can cost up to 65% less than equivalent public cloud storage. While such estimates are vendor-produced and should be independently validated for your specific workload mix, the directional argument — that public cloud storage can become expensive at scale — is broadly accepted in the industry.
The practical question Brown poses is a useful framing device for any architecture decision:
“The cloud has proven not to be as cheap as businesses were, perhaps, expecting it to be, but what’s the trade-off? Is converting ‘time-to-market’ faster more of a business focus than bottom-line operating costs are?”
Eran Brown, CTO, Infinidat EMEA
Designing a Hybrid Architecture That Works
Hybrid cloud delivers meaningful value only when the private and public components are chosen to complement each other. Key design considerations include:
- Basepoint selection: The private cloud platform you build on (VMware, OpenStack, Nutanix, or a hyperscaler’s on-premises offering such as AWS Outposts or Azure Stack) determines integration complexity and long-term portability.
- Networking: A reliable, low-latency WAN connection — or a dedicated interconnect such as AWS Direct Connect or Azure ExpressRoute — is a prerequisite for consistent performance across the hybrid boundary.
- Data tiering: Define clear policies for hot, warm, and cold data so that active workloads run closest to compute and archive data lands on the cheapest available tier.
- Security and compliance: Identity federation, encryption in transit and at rest, and audit logging must span both environments, not just the public cloud side.
- Avoid over-engineering: For many organisations, augmenting a public cloud with targeted private capacity is more pragmatic than building a fully symmetrical private platform. The goal is solving specific problems, not replicating the hyperscaler’s feature set on-premises.
Beyond Hybrid: The Multi-Cloud Direction
Alongside hybrid cloud, a growing number of enterprises are adopting multi-cloud strategies — deliberately distributing workloads across several public providers and private clouds. Multi-cloud is more operationally complex than a straightforward hybrid model, but it offers concrete benefits:
- Resilience: No single cloud provider becomes a single point of failure.
- Reduced DDoS exposure: Distributing services across providers limits the blast radius of targeted attacks.
- Cost optimisation: Workloads can be placed on the provider offering the best price/performance for that specific task.
- Vendor leverage: Avoiding lock-in preserves negotiating power at renewal time.
Brown sees multi-cloud as the natural evolution for organisations that have already validated a hybrid approach:
“Data gravity issues and digitalisation may result in increased interest in multi-cloud solutions which are better able to compete on price, availability and resiliency.”
Eran Brown, CTO, Infinidat EMEA
Cloud Strategy Defines What Stays On-Premises Too
One of the most overlooked aspects of any cloud strategy is that it is, by definition, also a decision about what does not move to the cloud. Regulatory requirements, data sovereignty laws, latency-sensitive workloads, and intellectual property concerns may all point toward retaining certain systems on-premises indefinitely. Brown’s closing observation captures this well:
“Hybrid cloud models will continue to provide the most commercially resilient solution, especially for those companies that want to take a ‘cloud-first’ approach. A subset of defining a cloud strategy is also to consider that, inadvertently, you are defining what doesn’t go across to the cloud.”
Eran Brown, CTO, Infinidat EMEA
Key Takeaways for DevOps and Infrastructure Teams
- Hybrid cloud is a deliberate architectural decision, not a default landing zone. Evaluate it against specific workload and data requirements.
- Public cloud storage costs grow non-linearly with data volume — model egress, API, and regional fees alongside the headline storage price.
- On-premises private cloud can reduce storage costs for high-volume workloads, but requires upfront capital and ongoing operational expertise.
- The strategic question is not public vs. private — it is which data needs which properties (latency, sovereignty, cost, availability) and which platform delivers them best.
- Multi-cloud is the next logical step for mature hybrid adopters, bringing greater resilience at the cost of increased orchestration complexity.
- Every cloud migration strategy implicitly defines a retention strategy. Make that decision explicitly.