The future of data warehousing in the IBM Cloud
Today, we reimagine data warehousing in the IBM Cloud.
We’re announcing a brand new service, a complete paradigm shift in our approach to Cloud data warehousing. It’s a service that finally delivers on the promise of Cloud computing: true elasticity coupled with seemingly unlimited resources, packaged as a fully managed offering. We’re calling this new addition to our family, Flex Performance. It’s the flagship tier of a brand new line of Flex offerings. It’s available exclusively through the IBM Cloud. And, we think you’re going to love it.
Here’s what we’re bringing to market today:
Independent scaling of storage and compute: Easily scale up (burst) the compute on your warehouse during peak demand, then scale it back down to steady state to reduce your costs. And, if you fill up your warehouse to capacity, more storage is a click away. In other words, you configure your elastic data warehouse to fit your exact business needs. With 440 compute and storage combinations, you’re guaranteed to find one that works for you.
Self-service backup & restore: Retain up to seven backups of your entire cluster. Quickly restore your cluster to a previous state, if needed. And, schedule your backups to align them with the beat of your business.
Fast-recovery HA: Our bare metal compute nodes are now managed by the IBM Container Service. Any unexpected node failure is immediately detected, the node taken out of commission, and a new one is brought online from a hot standby pool. All completely automated.
Now, let’s talk about the hardware.
Each data warehouse is deployed on dedicated, bare metal hardware running the latest 2.6GHz Intel Xeon (Broadwell) processors. The base cluster starts off with 48 (real) compute cores, 1TB of high-performance memory, and 2.4TB of storage for your table data. And, every 24-core compute increment adds 512GB of memory to your cluster. That entire capacity is dedicated solely to your SQL queries, and nothing else.
Next, let’s talk storage.
All of your data resides on a highly-performant, SSD-backed block store at 12 IOPS per GB. It’s lightning fast. You can easily add more storage to your cluster in 2.4TB increments. And, whenever you increase your storage, we automatically provision other storage tiers that provide space for backups, scratch and temp space. And, of course, your data is automatically compressed when it’s ingested into the engine so that the initial 2.4TB of storage can easily hold more than 6TB of raw table data.
We’re very proud of what we’ve built. I’m very proud of the hard work that our extraordinary engineering team has put into this project. We’ve built the future of data warehousing in the Cloud, a beautiful architecture, and a beautiful platform to launch all future innovation.
We’re excited. You should be excited. And you’re definitely going to want to give this bad boy a spin.