by Joe Gleinser
27. May 2010 18:20
Your intrepid reporter recently went behind the scenes at a national cloud hosting provider. Once I pulled back the curtain, would you believe I found:
1) Single points of failure at the firewall and switch level. Yep. For an organization that spent millions on infrastructure and management tools, why not configure your firewall and switches for High Availability. The devices they used all supported HA configurations.
2) No backup of static configurations. If one of the single-points did fail, surely they could recover quickly? Right? So where do they store the configuration backups of these devices? The staff I spoke to had no idea. Reconfiguring a shared firewall and switches from scratch would be no fun at all.
3) Lack of spare parts. Anyone would stock replacement gear if they designed in a single points of failure. Not this cloud provider. Though the staff asserted that spare switches and firewalls exist, we were unable to locate the equipment. No spare firewall. No spare switch. Supposing it does exist, the replacement process apparently starts like this:
"Step 1: Dig through mountain of parts hoping to find replacement gear."
I'm as optimistic about the cloud as anyone, but caveat emptor.