Get the highlights in your inbox every week.
Interview with John Dickinson of SwiftStack and OpenStack Swift
Making cloud storage easy with OpenStack Swift
When you want to learn about object storage in OpenStack, John Dickinson is the guy to ask. John is the Director of Technology at SwiftStack, a company which relies on the OpenStack Swift project to provide unstructured data storage to customers around the world. He also serves as the Program Technical Lead (PTL) for OpenStack Swift and has been involved in the development of Swift since 2009.
At next week's OpenStack Summit in Paris, John is giving three talks: a Swift 101 course for beginners, an introduction to building applications with Swift, and a talk on extending Swift through its modular interface.
John gave us a preview of storage policies back in February; we caught up with him to learn what features made it into the Juno release, how SwiftStack is succeeding with an open source business model, and what's ahead for OpenStack Swift in the Kilo release cycle.
What's new in OpenStack Swift for the Juno release?
There are a number of new and exciting features in the Juno release, including:
- Storage policies which you can learn more about below;
- Keystone v3 support for better interoperability across all OpenStack projects;
- Server-side account-to-account copy extends existing copy functionality in Swift to allow users to copy data between accounts without needing to first download it locally;
- Better partition placement when adding a new server, zone, or region so that adding capacity is easier for system admins;
- Zero-copy GET responses using splice makes storage servers more efficient; and
- Parallel object auditor allows for more efficient auditing and faster detection of bit-rot errors.
What are storage policies and what is exciting about them?
Storage policies allow you to tailor your storage infrastructure to exactly match your use case. They allow deployers to specifically configure their Swift cluster to support the different needs of data stored in the cluster, and they provide unrivaled freedom and flexibility for storage services their users and applications need. IT can deploy storage policies that segment across geographically dispersed data centers, different storage hardware performance or by number of replicas. Tiers of storage are consolidated into a single system, management becomes simpler, costs drop and IT bandwidth becomes more available.
How is SwiftStack able to build their business model successfully around the Swift project?
Swift is a really good storage engine, but when you deploy it in a brand new place, you need more information and pieces to integrate it into existing IT infrastructure. That’s where SwiftStack come in. Swift is the open source storage engine, and SwiftStack is the software that makes Swift simple to use and integrate with other systems.
Also, community is vitally important to us. We're thrilled to be a part of the OpenStack community as the leading developer for Swift, but we’re also working with huge blue-chip companies like HP, IBM, Intel, RedHat, and Seagate to future contribute to the open source community.
Looking to the future, what are you hoping to add in Kilo and beyond?
Kilo is the next OpenStack integration release, which will include erasure code support, Swift encryption, and other features to be announced. Stay tuned!
How is the object storage paradigm changing how developers write software in the cloud era?
More and more companies need a storage solution that can scale as quickly as the business grows. Object storage gives businesses an easy-to-use, scalable and cost-saving storage solution that legacy file-based systems can’t offer. In particular with OpenStack Swift, developers of software and hardware applications can focus on providing value in their app. The hard problems of using storage—like scale, availability, and failure handling—are automatically handled by Swift. For deployers, Swift allows for smooth capacity management, simple operations, and transparent upgrades. OpenStack Swift provides the storage engine that grows with you instead of inhibiting you.
What advice would you have for someone wanting to get started with Swift?
If you are starting to use Swift, there are questions that should be asked. What are you trying to accomplish? What issues do you currently face with your storage? Depending on the use case and what you want to accomplish with your storage system, Swift may be a great option for you. Also, take a look at the OpenStack Swift page on SwiftStack’s website. And if you have questions, feel free to stop by and ask questions in the #openstack-swift IRC channel on freenode.net.
Is there anything else you'd like to add?
Check out our sessions at OpenStack Paris next week. We’ll be discussing everything from Swift 101 to Storage Policies, use cases and how to build apps with Swift. Follow us @SwiftStack too for updates on news and our whereabouts!