The Cloud Is Built on Uniformity. Databases Are Not

When engineers build infrastructure, they aim for stability. Systems must behave predictably under load. They must survive success as well as failure. Growth is assumed. If the system is successful, it will be stressed. The challenge is ensuring that growth does not destabilise the structure meant to support it.

Uniform Supply, Irregular Demand

The public cloud addresses this challenge through uniformity. Compute, storage and network capacity are delivered as standardised building blocks. Instance sizes are predefined. Performance characteristics are abstracted. Capacity can be replicated across availability zones and regions with remarkable consistency. This is the operational achievement of Amazon Web Services, Microsoft Azure and Google Cloud. Uniform supply enables hyperscale.

It is an extraordinary achievement. It is also deliberately impersonal.

Enterprise systems of record operate under different conditions. They reflect the behaviour of real customers and real businesses. Demand is rarely smooth. It clusters around product launches, reporting cycles, market opens, regulatory deadlines and seasonal peaks. A transactional database sits at the centre of this behaviour. It absorbs whatever the business generates.

It does not get to choose the shape of that demand.

Cloud infrastructure assumes that workloads can be distributed horizontally across uniform units. Systems of record often experience concentrated pressure at specific moments in time. This is not a flaw, although it can feel like one at 9:01 on Monday when performance dashboards are flashing red. It reflects the difference between uniform supply and irregular demand.


Statistical Guarantees, Deterministic Commit

The distinction becomes clearer when examining reliability. Cloud platforms express resilience statistically. Availability is measured in percentages. Durability in strings of nines. Performance in ranges and envelopes. At fleet scale, these models are extraordinarily effective.

Most of the time, that is enough.

Transactional systems optimise for a different property. When a transaction commits, it must commit correctly and in order. When state changes, it must change once. The authoritative copy of truth cannot be approximate. Modern database engines distribute storage and parallelise reads. But the semantics of correctness still converge at the transaction boundary. Correctness is deterministic even when the infrastructure beneath it is statistical.

Cloud platforms optimise for aggregate behaviour across fleets. Transactional systems optimise for the one transaction that must not be wrong. Both approaches are rational. They are aimed at different targets — and those targets matter most under stress.


Where the Models Converge

The economic model reinforces the distinction. Elasticity allows capacity to expand and contract in response to demand. Stateless services align naturally with this pattern. Horizontal replication tends to make cost proportional to usage.

That is the promise.

Systems of record often scale differently. Reducing latency, increasing memory and tightening the coupling between compute and data frequently precede horizontal distribution. Cloud providers offer very large instance types and strong performance isolation. It is entirely possible to achieve impressive stability and scale in this way — and not infrequently, equally impressive cloud bills.

However, as performance guarantees increase, elasticity typically narrows. Dedicated capacity replaces shared pools. Provisioned throughput replaces best-effort scheduling. Predictability rises. Flexibility declines. These are trade-offs rather than failures. They always have been.

For many years, architects could separate these optimisation philosophies. Operational systems preserved deterministic state transitions. Analytical platforms absorbed variability downstream.


AI Removes the Insulation

AI reduces that separation. Machine agents increasingly interact directly with live operational data rather than curated extracts. Inference engines that are inherently probabilistic now depend more directly on infrastructure expected to be exact. The probabilistic layer now relies on deterministic state in real-time.

This does not invalidate the cloud model, nor does it imply that systems of record do not belong there. It makes the architectural assumptions more visible. Uniform, statistical infrastructure interacts with non-uniform, deterministic state, and the optimisation choices at each layer become more tightly coupled.

Architects must therefore decide explicitly where determinism resides in their stack and what elasticity they are willing to trade to preserve it. Elasticity, isolation and correctness can coexist. But not infinitely, and never without cost.

Someone always pays for certainty.

The cloud is built on uniformity. Databases are built on authority. That authority carries operational consequence.

As AI tightens the feedback loop between them, trade-offs once hidden inside abstraction become explicit design decisions. And explicit design decisions are where responsibility ultimately resides.

This article is part of the Databases in the Age of AI series.

Why We Spent 20 Years Protecting Databases from Analytics (and Why AI Just Broke That Truce)

8:53am on a Monday

I want to take you back roughly twenty years.

It is 8:53am on a Monday morning in central London and I am sitting at my desk staring at the screen. My coffee is untouched. My palms are sweating.

The ETL job is still running.

If that phrase means nothing to you, it was enough in the mid-2000s to strike fear into the heart of any production DBA. At the time I was the database lead for a SaaS platform serving customers across the UK, most of whom would start logging in from 9am.

The weekend ETL process (Extract, Transform, Load) pulled data out of the operational database, reshaped it and pushed it into the enterprise data warehouse. It ran every weekend and it was supposed to finish long before Monday morning.

It had not.

CPU pinned. Redo logs churning. User sessions beginning to queue.

In a few minutes customers would start calling support. Support would start calling me. And somewhere upstairs, the CEO would notice that his dashboard was still showing last week’s numbers.

I had two options. Let it run and hope it finished before the system buckled under real user traffic. Or kill it and spend the next hour watching rollback potentially take even longer, while guaranteeing that the warehouse would be stale until the following weekend.

How did we end up building systems where that was a normal Monday morning dilemma?


Why We Built the Wall

For the better part of two decades, the industry answered that question in one consistent way: keep analytics away from operational systems.

Transactional databases were designed to process orders, update accounts and record events predictably. Analytical workloads were different. They were heavy, exploratory and often poorly constrained. They scanned large portions of data, built aggregates, joined everything to everything and consumed CPU and I/O in bursts that were difficult to forecast.

Putting the two together in the same system was a recipe for contention.

So we separated them.

We built ETL pipelines. We built data warehouses. Later, we built data lakes and lakehouses. We introduced replication, change data capture and streaming. Each innovation was, in its own way, an attempt to preserve the integrity of the system of record while still making data available for analysis.

This was not fashion. It was defensive architecture.

The separation protected revenue-generating systems from analytical curiosity. It provided workload isolation. It gave operations teams a fighting chance of keeping Monday morning uneventful.


It Was Never Just About Performance

Enterprise environments rarely have a single system of record. A CRM system holds customer interactions. An ordering platform tracks transactions. Billing lives somewhere else. Supply chain somewhere else again.

The warehouse became not just a safety valve for analytics, but a unifying layer. It was the place where disparate operational systems could be reconciled into something coherent.

For years, this model worked.

Dashboards were allowed to be slightly stale. Reports could reflect yesterday’s state. Humans tolerate delay. In fact, they often prefer it. Analysis takes time, and decisions are rarely made in milliseconds.

The truce held because the consumer was human.


AI Changes the Consumer

AI agents change that.

An AI agent does not log in at 9am. It does not wait for a dashboard refresh. It does not tolerate yesterday’s numbers if it is expected to act on what is true right now.

Inference is not reporting. It is decision execution at machine speed.

If an agent is recommending a next action, approving a transaction, adjusting a price or triggering a workflow, the freshness of the underlying data becomes materially important. Close enough is no longer good enough. Staleness is no longer cosmetic. It alters outcomes.

The architectural assumption that analytics can safely run on a delayed copy of operational data begins to fracture.

This does not mean warehouses were a mistake. It does not mean lakehouses are obsolete. It does not mean streaming pipelines were misguided.

It means they were optimised for a different consumer.

For years, we optimised for human analysis. Now we are increasingly optimising for machine-driven action.

That is a different problem.


The Balance of Trade-Offs

For two decades, the answer was clear: keep them apart.

Protect the system of record. Move the data. Analyse it somewhere else. Accept a little delay in exchange for stability and control.

That architecture was forged in moments exactly like that Monday morning at 8:53am, CPU pinned, redo logs churning, business users about to log in.

AI does not invalidate that history. It simply changes the balance of trade-offs.

The truce between operational databases and analytics was built for a world where humans consumed insight.

We are now entering a world where machines consume state.

And that changes the conversation.

This article is part of the Databases in the Age of AI series.