One challenging aspect of enterprise data storage is that it relies on three different storage protocols:
- Block storage breaks data into fixed blocks with unique identifiers, and is typically used for databases and other workloads requiring high performance and low latency;
- File storage is used for storing spreadsheets, graphics, images, and other documents in hierarchical folders;
- Object storage is used for managing massive amounts of unstructured data, and is increasingly popular today for artificial intelligence (AI) and advanced analytical applications.
The protocol challenge has traditionally been solved with separate, specialized storage systems, but as the hybrid cloud operating model has become the de facto standard, savvy IT leaders are migrating to software-defined storage solutions that leverage open-source software development and industry-standard x86 hardware.
That’s where IBM Storage Ceph fits in. Storage Ceph consolidates support for block, file, and object storage protocols to help organizations eliminate data silos and deliver a cloud-like experience while retaining the cost benefits and data sovereignty advantages of on-premises IT.
Today we’re announcing Storage Ceph 7.1, which adds support for NVMe-over-TCP, delivering block storage capabilities to VMware environments over standard protocols that are transparent to the VMs. This will enable VMware users to decouple VMware storage from compute, offering more flexibility and potentially lower cost.
Storage Ceph is designed to deliver solid performance over cost-effective TCP networking.
It offers the flexibility of scale at any size – smaller environments can benefit from a single implementation with all three storage protocols, while larger organizations can scale out a single instance or choose to segment into various clusters while maintaining durability and cost-effectiveness.
In the broader context of IT strategy, the popularity of the on-premises hybrid cloud model has organizations looking to bridge the architectural divide between IT Ops and DevOps. Traditional enterprise apps relying on block storage systems running in on-premises data centers now need to work seamlessly with webscale apps in bare metal and containerized environments connecting to the cloud via platform-as-a-service. Storage Ceph is designed to span these different infrastructures, application architectures, and different management and consumption experiences.
Storage Ceph data pools can be accessed via block protocols (for structured data) and via AWS S3 compatible REST APIs (for unstructured data), making it ideal for both application storage and as back-end storage for data lakehouses. This makes it a good choice for organizations that are now shifting their AI deployments from proof-of-concept to full production and are looking to gain increased flexibility and lower costs by moving some cloud-based workloads to their on-premises data centers. Storage Ceph is also compatible with watsonx.data, the Snowflake data cloud, and Apache Spark.
Storage Ceph is designed to be deployed and managed with minimal time and effort, an important factor today given how costly and difficult it’s become to hire skilled IT staff, both generalists and specialists. The problem is particularly severe in organizations that migrated many applications to the cloud over the past decade and no longer have a large pool of skilled in-house IT personnel. Storage Ceph addresses this challenge with a scale-out architecture that is self-healing and self-managing, providing a highly elastic and reliable storage experience with minimal user intervention.
Other new capabilities in Storage Ceph 7.1 include:
- vCenter Remote plugin for VMware;
- Intel QAT acceleration for object check summing, encryption, and compression;
- NFS v3 and NFS v4 support;
- CephFS replication improvements and broadened functionality;
- CephFS snapshot management from the dashboard UI.
For organizations that want the benefit of on-premises software-defined storage without the hassles of integrating and operating the necessary hardware, IBM Storage Ceph Ready Nodes provide an integrated solution – pre-configured and pre-validated industry-standard servers with hardware and software support provided by IBM.
Storage Ceph Ready Nodes employ a simple building block approach to provide organizations with flexible configuration options that have been tested and certified with Storage Ceph software, so companies can start with a configuration that meets their current needs and simply scale with additional capacity when required.
Storage Ceph 7.1 will be generally available on June 14, 2024; Storage Ceph Ready Nodes are generally available now.
For more information, check out:
- IBM Storage Ceph Webpage;
- IBM Storage Ceph Data Sheet;
- IBM Storage Ceph Solutions Guide;
- IBM Storage Ceph Concepts and Architecture Guide;
- IBM Storage Ready Node Data Sheet.
#ibmtechxchange-ai