Azure Storage + Everpure: Managing the New 'Blast Radius' in AI-Ready Infrastructure

Описание: Protecting the Petabyte: Managing the New “Blast Radius” in AI-Ready Infrastructure | Ben di Qual (Everpure)

When your world moves from “big data” to petabyte reality, the rules change.

AI doesn’t just scale your data…
it scales the consequences of every decision you make about performance, locality, security, and recovery.

In this episode of Azure Storage Talk, I’m joined by Ben di Qual (Everpure formally Pure Storage) for a conversation that’s not a product pitch—it’s a real look at what breaks first at scale, why “infinite cloud storage” is often misunderstood, and how leaders can shrink the blast radius in AI-ready infrastructure.

We get into:

Why “limitless capacity + performance” still hits real ceilings
The locality gap (where compute lands vs where data can actually keep up)
Why training ≠ inference (and why treating them the same gets expensive fast)
What “AI-ready storage” should mean in practice (checkpointing, tiers, tradeoffs)
Why incidents turn into board-level moments—and what restores trust afterward
The practical “radius reducers”: boundaries, segmentation, immutability, recovery readiness

👇 QUESTION FOR YOU:
What’s your biggest petabyte-era worry right now: performance, locality, security, or recovery?

Chapters
00:00 Petabyte drives & the new “blast radius”
01:00 Welcome + why this is a “belief systems” conversation
02:00 The “infinite cloud storage” myth
03:00 Performance ceilings + the locality gap (GPU placement vs storage)
04:00 Lift-and-shift vs real cloud value
05:00 The missed step: refactoring + shared-nothing scale out
06:00 What “AI-ready storage” actually means
07:00 Checkpointing tiers + why old patterns are breaking
10:00 Marketing vs reality: training ≠ inference
11:00 The resource crunch (power, GPUs, NICs, lead times)
12:00 Pure + Azure: what’s exciting about the integration
13:00 After an incident: what leaders are really afraid of
16:00 Restoring trust: immutable snapshots + defense-in-depth
18:00 Balancing user autonomy with modern security risk
20:00 QLC + tradeoffs: doing more with less
21:00 Back to the petabyte: reducing blast radius at this scale
23:00 Power/footprint math + what happens when a server fails
24:00 What Ben wants you to check out next
25:00 Closing

#AzureStorage #AIInfrastructure #CloudStorage #DataPlatforms #CyberResilience #DisasterRecovery #PureStorage #MLOps