I'm really curious about the idea of multi-cloud as a potential solution after the recent AWS outage affecting DynamoDB. The Forbes article mentions how it could be a game-changer, but I want to hear from those who've actually worked with multi-cloud setups. Here are a few things I'm specifically interested in:
1. How quickly can you accurately detect an outage? They say it might be possible in about five minutes, but does that hold up when millions are on the line?
2. Is it feasible to replicate AWS DynamoDB to another cloud provider without running into issues with consistency and latency? I've noticed there's no straightforward way to do this with GCP/Azure/OCI.
3. What are the actual costs involved in setting up and maintaining a multi-cloud solution? Think about development, infrastructure, and potential revenue loss due to increased complexity.
4. Can a multi-cloud strategy really protect against every possible outage from a cloud service? Have you seen successful failovers?
5. Ultimately, does investing in a multi-cloud approach pay off in the long run? Are there any solid figures showing that it's worth the cost over several years?
It seems like most discussions around this topic are pretty generic. If you've built a multi-cloud solution, I'd love your insights!
1 Answer
When diving into a multi-cloud setup, I'd recommend avoiding vendor lock-in services. Instead, consider using something like MongoDB, which is cloud-agnostic, and manage your own replication and failover. This way, you won't be stuck dealing with each cloud provider's individual quirks. It’s a simpler approach for your entire stack too, like K8s and RDBMS.
That makes sense, but isn’t that kind of against the grain of what a managed cloud service is supposed to offer?