20 July, 2024
 |
| Delta Airlines airport transfer desk seen crowded due to airline cancellations and delays at Paris-Charles de Gaulle Airport today Photo/ Hector Vivas/Getty Images |
Today, the MS Azure cloud service experienced an outage, leaving hundreds or thousands of companies scrambling. What happened to all those businesses sharing the same Azure infrastructure?
Back in the '80s
and '90s, we gleefully dismantled mainframes, decrying them as bad because
everything was concentrated in one giant system. Thriving into this mad trend of the time, companies downsized en masse, adopting Unix-based client-server
systems for their IT needs.
Fast forward, and
we've flipped the script yet again. Now, all business applications must migrate
to colossal cloud services, essentially replicating the centralized mainframe
model under a shiny new guise. This single-company, single-infrastructure model
of cloud services is more precarious than any old mainframe.
Who sold us this
lunacy and why? All in the name of cost savings for IT operations and
maintenance, of course. But today’s outage exposes the fallacy of this logic:
sharing a massive cloud infrastructure managed by a single provider is more
perilous than the mainframes we once condemned.
Just as
offshoring manufacturing to China led to unforeseen dependencies, consolidating
IT infrastructure into a single cloud service provider has created gigantic vulnerabilities. Now, businesses find themselves at the mercy of one provider's
failures, with the fallout extending from Western corporations into their former
colonies. This invisible tethering underscores a lingering colonialist dynamic,
an unbroken line from past imperial glories to modern-day dependencies.
The world hasn't
evolved much in this regard; we've merely upgraded our excuses and
surface-level solutions. The inherent risks of these massive, shared
infrastructures remain, and today’s outage is a stark reminder of our folly.