It’s an unseasonably hot afternoon in Seattle. She’s leaning on the railing of the Pike Street overpass, psychedelic gypsy skirt, black tank top, Audrey Hepburn sunglasses. Her hair is a long, brunette ponytail. She has a partially-completed tattoo sleeve on her left shoulder. It was started a long time ago.
The cars stream down the highway, bright shafts of reflected sunlight seeking uncovered eyes. Hundreds of people on their way somewhere, swimming downstream with society at speed. She’s like a statue. I’m standing with a homeless man by his cardboard encampment. Neither of us are doing what we’re supposed to do right now.
But it can’t last forever.
‘Spare some change?’
I look over, and hand the homeless man $10.
Of course, she’s gone by the time I turn back. Even my memories of her start to slip away, again. True memories, like fish, are deliberately slippery. They’re hard to keep hold of at the best of times.
A quick note, following from news of Google Compute outage yesterday, and outages caused by DNS changes at Amazon S3 slightly more than a week ago, it’s important to remember that moving to the cloud still requires operations (sysadmin, devops, whatever we want to call it).
There is a belief that moving to the public cloud allows companies to outsource most, if not all of their operations staff. But there is a very real danger in abrogation of ops responsibility.
If you are outsourced to the cloud, do you have a disaster recovery plan? What happens if the systems that are ‘too big to fail’ do just that?
I’m not saying the cloud is bad — it enables companies to go to the web with next to no capex investment. But that doesn’t mean it is the end all be all, and if you’re not taking care of operations in-house, it’s very likely that you will regret it.
Here’s some further interesting reading: TechTarget Cloud outage report 2014.