Last week, we disappointed everyone who thought sticking to one cloud provider would make their lives easier and less expensive. This week, we’re back to disappoint you again. Here’s why the other option, operating across clouds, is just as frustrating…with one major exception.

But before we get there, let’s talk about some of the benefits of operating across multiple clouds. For one thing, having multiple cloud providers courting you means you gain the power to negotiate your agreements. You can shop multiple offers, and if Azure raises their prices, you can threaten to leave them for AWS. Plus, having access to multiple cloud environments gives your teams flexibility and permission to use the tools they like. It stops the “shadow IT” rebellion before it starts.

Freedom and flexibility are nothing to sneeze at, but I did say there was more disappointment coming, and here it is: operating across multiple clouds is a strong invitation to bloat. Once a cloud provider’s sales team gets 3% of your spend, they’ll try to get you to increase that share to 30% (recall the “land and expand” approach). They’re fishermen using barbed hooks to catch their dinner - not to release you back into the sea, where you can explore other products outside their proprietary, restrictive tools.

Once you’ve taken the bait and are hooked up on their barb, you’ll have a tough time shaking free. So what do you do? You buy even more of their tools, until you’re living in a duplicative, restrictive world, where negotiation isn’t the game anymore. At this point, you’re inviting your teams to become disorganized and siloed.

“But wait!” you say. “There are other choices besides those, and even the big cloud providers have open source options. I’m not stuck, right?” Yes and no. Open source platforms like Postgres, Kafka, and Apache Spark are cloud-neutral, which means that in theory you can avoid the trap of duplicative, restrictive tools. But in spite of this inviting premise, the landscape is still difficult to navigate and isn’t nearly as “open” as they’d like you to believe.

You see, many of these open source companies have a more subtle barb: they’re open core, not truly open source. Even though they run on an open source core, and are therefore ostensibly more flexible in nature, they still incorporate a growing collection of features to differentiate from their own open source, all of which make it difficult to leave. It’s understandable for companies to go this route. It’s hard to make money with an undifferentiated commercial product. So any open source company that doesn’t want to engage in a race to the bottom may do what they have to do.

So what have we learned? Operating across clouds isn’t great…but it’s better than the alternative. This strategy can be just as confusing as managing one cloud provider. But if you do it right, it can be beneficial.

I did say there was one major exception, and I’ve already kind of introduced it to you. New barbless, open source models could change the game. Their primary aim is offering a better user/developer experience.

While open source is an answer to the bloated data stack, the market still has problems. It is difficult to navigate and many leaders are moving towards restrictive commercialization vs. true open source. Tune in next week to hear about how Tembo is bringing clarity to the market as a more pure open source solution (and still building a strong business in the process).

Ry’s Weekly Resources