Creating scalable, intelligent storage solutions with OpenStack
Managing complexity and the sheer volume of storage requirements within the corporate environment today is one of the greatest challenges facing IT departments. The growth of business data and the insatiable demand for storage has been a catalyst for developing a new approach to enterprise storage in the cloud.
In an earlier article, we examined how software-defined storage (SDS) can lower cost and increase storage efficiency. In this article, we'll explore how an Intel-powered OpenStack cloud can provide scalable, intelligent storage solutions to meet growing business needs.
Data stored within the enterprise can generally be classified into two distinct groups: Mission-critical or "hot" data, which needs to be accessed rapidly or modified frequently; and larger, less frequently accessed or "cold" data sets that often need to be retained for long time periods, due to regulatory requirements.
The notion of tiered storage is an important concept in the software-defined storage model. Data is assigned into predefined categories based on specific criteria such as performance requirements, frequency of access, or level of protection required. Each category is then mapped to specific types of storage media according to requirements. The "hot" storage tier requires high-performance, low-latency solutions, while the "cold" storage tier tolerates higher latency and can be stored on slower, cheaper, larger-capacity drives. Finding the right balance between performance and cost while ensuring availability can be challenging, but OpenStack cloud-based storage can help reduce overhead and optimise storage management.
Swift is the central component underpinning OpenStack's software-defined storage model. It implements a fully redundant, distributed API-driven storage platform using clusters of standard Intel servers. Swift also facilitates elastic vertical and horizontal data scaling, allowing virtually unlimited quantities of tiered storage to be added to the cloud as or when needed. To maintain integrity and protect from failures, Swift also features automated self-healing, auditing, and replication capabilities commonly found in other enterprise storage platforms.
Because Swift controls data replication and distribution across storage nodes in an OpenStack cloud using software-defined policies, commodity hard drives and servers can be used to keep costs low.
Intel Xeon processors and networking components provide high I/O throughput to deliver optimised performance necessary in compute-intensive hot storage solutions. Such solutions support powerful data protection mechanisms such as triple replication, erasure code, and data encryption. Storage acceleration applications that implement data deduplication, thin provisioning, and data compression further reduce costs by decreasing capacity requirements.
Intel Atom processors are ideal for cloud-based cold storage. Specifically designed for higher storage density and lower power consumption, Atom-powered servers provide a scalable, cost-effective archival and backup solution.
A typical enterprise cloud storage solution will incorporate an appropriate mix of Intel Xeon and Atom-based servers implementing both hot and cold storage. Red Hat Enterprise Linux OpenStack Platform and Swift can be used to unify storage of these tiers under a single control point, providing a powerful, scalable, software-defined storage solution capable of meeting the growing storage needs of any business.