"In today's Busting Big Data Myth's video, Tom Fisher, CTO at MapR, busts the myth of the cloud being a data platform. (Hint – it's not.)
Here's what Tom has found:
1. People go to the cloud because there’s little or no infrastructure management or set up.
2. However, you still have to choose all the elements you want to use and then assemble it.
3. If you have existing apps, you can containerize them and move them to the cloud, but the big issue is getting to your stateful data that’s still on-prem.
The complications grow as you do more.
Here are the top 5 issues:
1. Creating seamless on-prem to in-cloud is still hard, and data movement issues tie people up.
2. You can’t actually do analytics and AI/ML together. You still need separate cluster instances and have to stitch things together yourself.
3. There’s no common security model even within one cloud – let alone multiple clouds.
4. Each cloud is different and switching from cloud to cloud isn’t easy.
5. Edge-first architecture doesn’t really work in the cloud.
You can actually “Do Cloud Better” (and likely cheaper) by using a data platform inside a cloud. When you use a data platform, you get:
* 3x dev productivity because you don’t have to stitch it together yourself.
* And you don’t have to keep updating all the revs.
* You can create more powerful apps that combine analytics with AI/ML.
* Data pipelines and microservices implementations are simplified.
* You can bring data from edge to cloud reliably, securely, and scalably.
* Lower storage costs by automatic tiering into object store.
* And lastly: once you have based your cloud solution on a complete data platform, you can easily move the entire stack straight into a different cloud!"