Red Hat 블로그
By Daniel Gilfix, Red Hat Storage
This week’s announcement of Red Hat Ceph Storage 2 marks the most significant release of the cloud-native, software-defined storage product since the acquisition of Inktank by Red Hat in 2014. It represents an important milestone, not only in terms of the company’s steadfast commitment to storage but also from the perspective of preparing open source customers for the highly coveted software-defined datacenter.
Source: Scott Maxwell, LuMaxArt Golden Guy Trophy Winner, at https://www.flickr.com/photos/lumaxart/2293239853
Why storage matters
We work in an era where storage is often taken for granted, under-glamorized for the role it serves, and yet increasingly essential, often indispensable, to solutions spanning physical, virtual, private cloud, container, and public cloud environments. Nevertheless, in recent studies commissioned by Red Hat, Vanson Bourne Ltd reports that 71% of IT decision makers fear their organizations' storage solutions won’t be able to handle next-generation workloads, while 451 Research indicates that 57% already have or are moving to software-defined datacenters this year. As most loyal Ceph followers know, this is precisely why folks are so excited about the launch of our new product.
Source: Pixabay on public domain at https://pixabay.com/en/community-friends-globe-continents-909149/
What’s new in 2
While Red Hat Ceph Storage has distinguished itself as a unified storage platform that’s overwhelmingly preferred for OpenStack, for years Ceph has actually fulfilled this role with folks like service providers and the telco community as an object store proven at scale. Red Hat Ceph Storage 2 adds innovation resulting in a far more robust object storage solution for a wide variety of use cases aimed at the enterprise, like active archive, big data, and media content hosting. Customers also receive an easier-to-use product and life-cycle management by virtue of the integrated Red Hat Storage Console 2.
Object storage enhancements
Ceph’s object storage interface, RGW, has greatly improved. RGW geo-clusters with single namespace allows local communications with local clusters with “eventually consistent” synchronization between them. RBD mirroring enables multi-site replication for disaster recovery and archival. Support for Active Directory, LDAP, and Keystone v3 authentication systems expands security options. And improved S3 and Swift protocol API support--such as AWS v4 client signatures, bulk delete, and object versioning--strengthens compatibility with AWS and OpenStack, respectively.
In addition, there are three sets of functionality in Tech Preview for early customer exposure:
- CephFS, the POSIX-compliant file system that uses a Ceph storage cluster to store data in accordance with the OpenStack Manila service
- An NFS-to-S3 gateway for import and export of object data
- BlueStore, a new storage media backend that we expect to lead to 2-3X performance improvements for the entire product
All these features are instrumental to empowering Ceph to keep pace with the ever-growing demands of its spirited user base and to handle multi-petabyte workloads with the grace and efficiency that enterprise customers need for software-defined datacenters. On the surface, they might not appear especially sexy, but for cloud builders and IT decision makers of all sorts, many of whom are already in the loyal Ceph community, they are a breath of fresh air to an otherwise stifled march toward storage infrastructure agility.
See you at Summit
If you’re coming to Red Hat Summit at the Moscone Center in San Francisco between June 27-30, please stop by the Red Hat Storage booth (#31-32) or the partner pavilion (#529), where a cross-sample of Ceph partners are demonstrating the elements of our extended ecosystem. Further information is available as well on redhat.com/storage.