In the second post in my series, I outlined several possible storage architectures for systems that need to store both files and objects. In this post, I’ll discuss the design choice we made for Pure Storage® FlashBlade®.
First, here are a few points to think about when making an important architectural decision like this one:
When originally designing FlashBlade, we chose the file and object architecture that contains a file system and an object store as separate namespaces. It currently looks something like this:
We made this decision based on the vision that data centers will increasingly combine file and object workloads. We wanted to have a no-compromise file system and a no-compromise object store in one system, even if it would take longer to deliver certain features like cross-protocol access. While there are file/object cross-protocol access use cases we don’t support today, we’ll flesh out more of those over time.
FlashBlade provides a very solid foundation for both file and object workloads in a variety of real-world environments. Over time, we anticipate most, if not all, of our customers will leverage FlashBlade to unify their file and object workloads. Here are just a few examples:
These customers—and many others—have been able to consolidate multiple file and object workloads on the FlashBlade platform. As a result, they benefit from reliability and performance, while simplifying their environment and saving on operational costs.
It’s exciting to see the impact our system is having today, given that we intentionally made a bold, major architectural decision to build a storage system that would require some short-term sacrifice to get it right for the long term. But we wouldn’t have it any other way.
Design choices early on in a technology’s life cycle have a big impact on the long-term success of any system, so it’s important to choose well, and wisely.
Learn more about FlashBlade and how it can revolutionize your file and object workloads.