Red Hat Ceph Storage 2 Briefing Note
Software Defined Storage (SDS) continues to attract the attention of IT professionals, but broad adoption is not meeting the expectations IT professionals. Part of the challenge is that while SDS, in theory, breaks the chains of hardware lock-in, it does little to prevent software lock-in. A potential answer can be found in Open SDS but these solutions need to mature to deliver a package that is more approachable to the enterprise and cloud providers.
Ceph and Gluster are two of the most popular SDS solutions on the market. These are independent community projects, and are available under public licenses for maximum openness. Red Hat, of course, is a heavy investor in both projects, allowing it to deliver enterprise grade versions of both software applications, known as Red Hat Ceph Storage and Red Hat Gluster Storage, respectively. As we discussed in our Product Analysis “Open, Software Defined Storage – Ceph or Gluster?” there are areas where the two products overlap, and Red Hat has been busy trying to be more specific about what use cases are best for each product.
Defining the Ceph Use Case
OpenStack is an ideal workload for Ceph. Ceph is a massively distributed storage environment. Every piece of data that is stored into Ceph is divided into segments and then distributed across the nodes in the cluster. This highly parallel, high distributed storage method is tailor made for OpenStack, which itself distributes workloads.
If OpenStack was Ceph’s first “hit single”, then Red Hat wants its next hit to be object storage. Object storage is one of the fastest growing segments of the storage market and Red Hat Ceph storage is open and already being deployed as an object store at massive scale. Two attributes that should be well received by the typical object storage customer.
Red Hat Ceph Storage 2
At the heart of the Ceph storage software is an object storage foundation and in this release Red Hat chose to focus on bolstering its object storage capabilities to enable it to better compete in this market and to improve its attraction to more traditional enterprises. Part of that attraction is interfacing more directly with the enterprise infrastructure. In Red Hat Ceph 2 there is now support for LDAP and Active Directory and there is a tech preview of a new NFS gateway for data import and export.
Red Hat is not ignoring the cloud and service provider markets either. First they have increased their compatibility with the Amazon S3 API. Also in version 2 the solution provides global object storage clusters with a single namespace. This capability essentially brings a eventually consistent multi-site cluster capability to Ceph. Global clusters allow users to read and write from the cluster they are closest to at that moment in time. The clusters are strongly consistent within themselves and eventually consistent between locations, providing the appropriate balance of performance, consistency and data protection.
For block devices, replication is far more automated than it was in previous versions. In the past an administrator could pipe a snapshot through rsync and end up with remote replication capabilities, but this was obviously manual and required some expertise. In Red Hat Storage 2 block replication is synced automatically and is much easier to setup and operate.
Finally, Red Hat Ceph Storage 2 has a new interface, the Red Hat Storage Console. The new console makes it easier to add nodes as well as grow and shrink consoles, making managing the cluster life-cycle much easier to manage.
After substantial success in the OpenStack market, Red Hat is getting Ceph ready for its next big hit, the object storage market. The updates in version 2 clearly position Ceph for the object storage move. The new features not only and should make it more attractive to both enterprise and cloud provider markets for traditional object storage use cases it may also expand the object storage use case.