,
26 tweets,
4 min read
Read on Twitter
So people are piling on to Keith for daring to suggest that multi-cloud may not be a winning strategy.
Allow me to chime in and outright state that multi-cloud for a generic workload is, absent a compelling business reason, a moronic default.
Allow me to chime in and outright state that multi-cloud for a generic workload is, absent a compelling business reason, a moronic default.
"But you can save 20% on the compute cost of your containers" and pay many times that in transfer charges because data gravity always wins.
"If AWS goes under" then neither of us are having this conversation because we'll be making many tens of millions a year as consultants helping larger companies with much more expensive problems out of that mess.
"We can negotiate better rates" ignores that discounts are volume-tied, and "threatening to leave" is the sign of an inexperienced negotiator in this space.
"If Amazon becomes our competitor!" What do you mean 'if?' I don't *think* they're going to enter the "sarcastic self-mocking newsletter" space, but I'd not bet the farm on it. Their product strategy is "yes." Make your peace with that.
"We want to use various providers' free credits!" Yes. This is the mark of the serious businessperson. Let's go celebrate over a dinner in the sample area of Costco.
"We may need to move in five years" so consider it then. Building for multi-cloud today means you're giving up feature velocity in return for an option you may very well not survive long enough to exercise.
"We just signed a CorpDev deal with another provider" so it's probably time to migrate wholesale. I'm on board, but make for damned sure that the deal is worth it.
"Eventually you'll want to move everything to your own datacenter." Yeah, I like reading Hacker News fan fiction too, but some of us are trying to run businesses here.
"We acquired a division / company that's on another provider" and you should probably leave them there. Rule of thumb, the "single cloud provider" boundary is probably at the P&L.
"Multi-cloud is a very important thing to emphasize because--" if we go all-in on one provider you're going to have nothing you or your employer can effectively sell me? Just a guess there...
"But our Kubernetes Strategy" remains an oxymoron.
"Our primary provider sucks at X." Well, X better be pretty goddamned painful, because multi-provider integration is a barefoot walk in the park after it's been strewn with glass and set on fire.
"AWS growth has stalled so we want to hedge our bets--" I'll start to entertain this as soon as a second cloud provider breaks out cloud revenues explicitly. Until then it's a creative writing exercise for analysts.
"Our customers are all multi-cloud!" Yeah, and so am I unless I want to migrate off of G-Suite and close my GitHub account. Yes, both of those are in their company's respective cloud revenue buckets.
"No really, our customers are on the actual cloud platforms of several providers; they're True Multi-Cloud!" I'll accept that as soon as the second place provider's spend level exceeds 10% of the customer's total cloud spend.
"One of our engineers went rogue and spun something up in another provider--" is a problem that starts with your second employee. It's always going to happen. "We're not an AWS customer!" proclaim companies with over 90 AWS accounts tied to their corporate email domain.
"Well what about the wildly successful multi-cloud strategy that was announced and demonstrated on stage by..." Please finish that sentence. I'm not saying unicorns don't exist, but there sure aren't any in my yard.
"Our customers won't let us put their data on AWS." I feel your pain. That's a compelling business reason. If you can't negotiate it away, that deployment's going elsewhere. Make sure they pay an appropriate rate.
"You'll pay through at least one orifice for SLA violations!" If you tie your SLA to a single provider in a single region, you've got bigger problems. Show me a single AWS service global outage, then come talk to me. Draft better contracts.
"You're just saying that because you're trying to sell--" what, exactly? I've advised customers to migrate specific workloads off of AWS. I don't have a vested interest here.
"You're just an AWS shill." Nope. Not a partner, not a shill. Show me an Azure customer or two with a nine figure annual spend problem and I'll be there underwear-outside-the-pants fast. I go where the expensive problems are; don't mistake that for being a partisan.
"Our internal corporate culture demands" what amounts to a blood sacrifice? Your culture is your business, but don't pretend a strategic version of a suicide pact is somehow a broader industry best practice.
"But Pinterest / Netflix / Snap" Wait--don't you work at a credit union in Kansas? Maybe, and I'm just spitballing here, the scope and scale of your problems might be a smidgen different than theirs?
"You've gotta have a monetary incentive here." Use your head. Do you think I could make more money doing what I do now (fixing AWS bills) or steering enterprises into 5-year migration boondoggles?
That's a reasonable first pass attempt. Fight me! What did I miss, other than how YOUR employer has a magic service or product that makes everything I just talked about go away once people pay you?
