I’ve just finished a paper for NetApp about Object Storage and StorageGRID (here’s the link if you are interested in knowing more) and I wanted to share here the Executive Summary of the document to give you an idea about the content you’ll find in it:
Business and end users are asking much more of IT now. Mobility of people and devices, new data access patterns, modern applications, and IT consumerization are causing new needs. Users want the best experience, unconstrained access to their data and flexibility while, at the same time, continuous data growth, security concerns and rigidity of traditional IT infrastructures make it very difficult to respond adequately.
It’s now evident that cloud computing, supported by a modern storage infrastructure, is fundamental to coping with this situation. Technology is finally maturing and many organizations are now leaning towards hybrid cloud infrastructures, where they find the agility, flexibility and savings they need to support users and business processes. Leveraging private and public cloud at the same time, makes it possible to take full advantage of the reaction speed allowed by public cloud and the savings enabled by private infrastructures. Looking at this landscape, choosing the right hybrid multi-tier cloud storage is fundamental to manage data locality, meet regulations and company policies while providing tools to manage workload spikes and fast deployment of new projects.
Object storage is a cornerstone for a modern data storage strategy and can be used as a horizontal platform capable of serving many different needs including hybrid cloud storage deployments. Unmatched scalability, standard HTTP-based protocols plus the ability to geographically span across various locations are its main differentiators and enablers. It can also be considered the perfect repository for any kind of data that has to be stored for a long time (like logs, for example) and computed at a later time if needed (Big Data analytics).
NetApp StorageGRID has all the characteristics to be leveraged as a crucial building block in these types of scenarios for both enterprises and xSPs. A rich feature set and flexible deployment options (i.e. software-only or appliance) make it suitable for many applications and use cases, including data archives, media repositories and web applications data. Last but not least, its potential is amplified by the ability to leverage different data protection schemes (i.e. local and global erasure coding as well as multiple object replicas), Amazon S3 compatibility and cloud tiering capabilities (useful to manage at best spikes while maintaining cost under control).
Want to know more? here the link to download the paper.
Well, NetApp has struggled trying to transition from legacy NAS solutions to focus on object-based storage. Their new version of StorageGRID, moving their software to bare metal, instead of VMs which required VMware licensing, and adding S3 API compatibility are improvements. NetApp understands that exposing everything they do in storage via APIs is the right approach. That said, their S3-compatible API is not native. Only Cloudian has a native S3 API that is fully 100 percent compatible with the AWS S3 API. When considering the size of the AWS S3 ecosystem of solutions, only Cloudian can actually guarantee that all of them will work with Cloudian HyperStore. Everyone else, including NetApp, is going to have a hard time making that claim. In the end, object-based storage is only as useful as the size of the ecosystem of solutions that supports it. AWS enjoys tremendous network effects from its S3 storage, and every new storage solution that supports S3 makes AWS S3 more valuable to customers and fully S3-compatible storage vendors like Cloudian.
Hi Tim, just to correct you slightly, the StorageGRID S3 API is 100% native. There’s no protocol translation layer in there at all. I’d be interested to learn what makes you think that’s the case?
Cheers,
Ed
hi Ed, well there aren’t many OBS software vendors that use AWS S3 as their native API, so I was surprised to hear that StorageGRID uses S3 as its native API. This is different than claiming support for the AWS S3 API while using a non-S3 native API. Support for the AWS S3 API among OBS software vendors comes in degrees. There are 51 operations that you can perform in AWS S3. Nine of them are simple, 18 of them are considered intermediate, and 24 are advanced. Cloudian supports all of them. Cloudian also guarantees that any third party solution that is written to work with AWS S3 will work with a Cloudian cluster. Translating from a non-S3 native API to an S3 API would also introduce performance considerations, which based on your comment, is apparently not the case for StorageGRID. As to what makes me think that’s the case…I have a colleague who worked on the technical product marketing side of the house for 10 years at NetApp.