Hyperscalers have solved cloud infrastructure for most workloads. High-performance RDBMS remains the exception – the gap between cloud promises and database reality is still real.
The Biggest Gap In The Clouds? High Performance RDBMS
Category: Storage
Storage is the foundation on which database performance is built. Latency, throughput, and consistency at this layer directly shape how systems behave under load.
While often abstracted away in modern architectures, storage characteristics continue to define the limits of what databases can achieve – especially as workloads become more demanding and less predictable.
Cloud Compromises: Constrained and Optimized CPUs
Not all cloud vCPUs are equal. This article explains how constrained and optimised CPU configurations affect enterprise database performance in ways the spec sheet doesn’t reveal.
Overprovisioning: The Curse Of The Cloud
On-premises databases were overprovisioned for performance. Cloud databases are overprovisioned for a different reason: unpredictable pricing. This article explains the real cost.
The Public Cloud: The Hotel For Your Applications
The public cloud is like a hotel – flexible and convenient but shared infrastructure. This article explains what that model really means for enterprise databases.
Don’t Call It A Comeback
flashdba returns from retirement to ask the question nobody in the cloud conversation was asking: when 75% of databases move to the cloud, what actually happens to performance?
The Final Post: Hardware Is Dead
Seven years in the all-flash storage industry ends with one conclusion: hardware has lost the battle. The real competition now is in software, cloud and the economics of abstraction.
Flash Debrief: The End (part 1)
A retrospective on seven years in the All-Flash storage industry – from Violin Memory’s $2bn valuation to Chapter 11, through acquisitions and consolidation, to the conclusion that the flash wars had only one winner: cloud.
Oracle ASM and Thin Provisioning – How To Reclaim Space
Thin provisioning lets storage arrays report more capacity than they physically have – but Oracle ASM doesn't automatically return space when data is deleted. Here's how to reclaim it.
All Flash Arrays: Scale Up vs Scale Out (Part 2)
Scale-out adds performance by adding controller pairs, but forces unnecessary hardware spend when only capacity is needed. Combining independent scale-up and scale-out is the architecture that makes composable storage possible.
All Flash Arrays: Scale Up vs Scale Out (Part 1)
Enterprise storage has two independent requirements: capacity and performance. Scale-up-only architectures struggle to address both – hitting metadata and DRAM ceilings before either dimension is adequately satisfied.
flashdba 