Pure Storage has recently made an interesting announcement about a new set of cloud products and services designed to help its customers to manage hybrid-cloud infrastructures.
On one hand, there is Cloud Block Store based on Purity OS, the OS that runs their FA arrays. It is a virtual appliance in the form of an EC2 instance that has a set of functionalities similar to the physical arrays. On the other hand, they are extending their Snapshot functionality to S3 repositories with CloudSnap, opening up a lot of opportunities for long time snapshot retention, data protection, migrations, DR and so on.
These two products together enable Pure’s customers to implement better architectures based on a product and procedure they already know and trust. At the same time, thanks to the optimizations introduced, the overall TCO can be improved as well.
There is also a third part to this announcement, StorReduce, I’ll be talking about it in an upcoming post.
Hybrid Clouds Need Hybrid Ecosystems
In the last few years we have seen a lot of organizations embracing cloud computing. After an initial moment of skepticism some have jumped on the bandwagon like there was no tomorrow. Others, even if excited by the potential of this new computing paradigm, have moved slowly and have opted for a more conservative approach. After some time, we can all agree that an IT strategy based only on public cloud infrastructures is not feasible for the vast majority of organizations.
The Cloud brings a lot of agility and flexibility to the game, and everybody loves to pay for what is actually consumed, and granularly, instead of making large investments with all the risks associated with it. The flipside is cost, or its unpredictability at times, and this is why we have all this back-and-forth between on premises infrastructure and public cloud.
Hybrid- and Multi-cloud strategies are very common now, data management and movement across clouds is the bigger challenge.
Pure Data Cloud Data Services
With Cloud Data Services, Pure is bridging the gap between on-premises and cloud infrastructures. Pure’s customers will be able to take advantage of the same front-end functionalities and management features on different infrastructure backends, allowing to minimize the cost of data movement, especially egress costs, which are applied when you want to access data remotely.
For example, Cloud Block Store provides the same data footprint optimization available on Pure Storage arrays and this is also reflected on features like remote replication. Because of the limitations imposed by the type of deployment (a single AWS AMI instance), a single Cloud Block Store can neither reach the same level of availability nor the performance of a physical array, but Pure has other options if that’s what you’re looking for. What’s really cool is that Cloud Block Store can be deployed and clustered across two availability zones, making it a pretty solid solution for some applications that need strict HA requirements. Other use cases include remote replication between on-prem and cloud for DR or data migrations but, in my opinion, dev and test could be the other interesting use cases.
Cloud Snap, the second part of this announcement, is another piece that helps to complete the puzzle. The mechanism is really simple, yet powerful. CloudSnap is a feature of a Pure Storage all-flash array to offload snapshots to a less expensive repository. It’s good for long-term snapshot retentions, backups, and DR as well as data migrations. Secondary repositories could be FlashBlade or any other NFS and S3 target. This feature was already available, but has been expanded to include more than just FB and NFS. The snapshots are self-consistent, incremental, and space optimized. This mechanism is perfect for several use cases which include simple and automated backup operations, disaster recovery, long term data retention and data migrations.
As you can see, some of the use cases are really similar to what you can get from Cloud Block Store, but the characteristics of the product are different in terms of performance, speed of access and cost. The two solutions do not compete with each other and have a minimal overlap, leaving the customers with a lot of freedom when it’s time to design their hybrid cloud infrastructure and select the best option for their use cases.
CloudSnap can work with any S3 compatible storage system but Cloud Block Store is certified only to work with AWS at the moment. I don’t think there are any restrictions to get it on other major cloud providers, and AWS has the largest market share. But I hope to see Cloud Block Store available on Azure and GCP very soon, and it wouldn’t be a bad idea making Cloud Snap compatible with Azure Blog or GCP Cloud Storage. That would be brilliant and would make this solution much more compelling for end users with multi-cloud infrastructures.
Closing the circle
Pure has always been a good example of the good mix every storage vendor should have: innovation, sensible solutions, good sales model and compelling services. That’s what makes happy (returning) customers. I think this announcement introduces an interesting set of solutions that are fairly aligned with Pure’s overall strategy and philosophy and will keep customers confident about the company’s roadmap.
With the increasing demand of hybrid- and multi-cloud solutions it is very important to extend the product line-up to match customer needs. All vendors are racing to expand their offer in a similar way and I think it is really important to signal this intent to existing and potential customers as well as to the rest of the market.