With so many different storage options today, users have access to exciting and powerful storage technologies previously reserved for the multimillion-dollar enterprise elite. Finding the right RAID solution can often prove to be a lengthy and dizzying process entailing months of research, conference calls to multiple storage vendors and endless budget meetings with management to justify the new investment. Making the wrong buying decision can often result in catastrophic data loss, unforeseen bandwidth bottlenecks, loss of man hours and precious company resources wasted on solving technical issues.
RAID storage usually makes most sense for users whose data plays a vital role in their operation. As storage systems develop and grow, more users find the need for large amounts of data to be stored and made accessible for future use. Combined with the rapid price drops for disk media, storage capacity and security have never been more affordable. The basic reasons for implementing a RAID array still remain the same: It provides a large storage capacity for a specific purpose or need, it complements and/or enhances system or application performance, and it protects and makes stored data easily available.
Internal PCI RAID storage
With the lowest cost, internal PCI RAID controllers remain a popular choice when compared with external RAID systems. PCI RAID entails a low deployment cost (ranging from several hundred dollars up to several thousand), delivers excellent performance and can offer high I/O capabilities. However, the use of internal PCI RAID controllers creates a different set of problems when it comes to maintenance and availability of the stored data. Because there is no such thing as a “one size fits all” PCI RAID system, critical component decisions are often left to the user. This results in the selection of components that may reduce the overall integrity and performance of the RAID array — the very reason why the system was deployed in the first place. By using an internal PCI RAID controller, the storage becomes a part of the host computer system and eliminates the ability to perform quick system maintenance, while lowering data availability and increasing downtime. Choosing a prepackaged PCI RAID solution may offer better overall compatibility and support in the event of failure or disaster.
External host-independent RAID storage
External RAID offers the greatest flexibility, integrity and deployment options compared with PCI and server-based RAID products. A host-independent RAID array is a storage appliance that has been designed with a specific purpose in mind: store and protect valuable data. Every component has been developed to perform a specific function, which increases the overall efficiency and reliability of the system. This type of system is normally classified as an “enterprise” product due to its increased integrity, redundancy and appliance-type nature. Host-independent RAID storage exists autonomously and does not require a dedicated host server; in fact, most of today's external RAID systems offer some form of high-speed network connectivity, eliminating the need for a host computer to be physically connected. Another advantage of external systems is the ability to dynamically repurpose or relocate the storage as needs change. Common user tasks can be streamlined or automated through the use of embedded or add-on software, which provides users with powerful management tools such as array expansion, volume snapshots, volume and file sharing, data replication/duplication to remote locations, automated backup, storage virtualization, and data life cycle management.
External server-based RAID storage
Offering a mix of low-cost, external RAID benefits and PCI RAID performance, server-based RAID storage is quickly growing in popularity both in the external RAID and SAN/NAS markets. Using common, off-the-shelf components (motherboard, CPU, memory, PCI RAID controller, etc.), server-based RAID storage can greatly reduce the overall purchase price compared with external RAID systems and typically offers some form of expansion options for future growth.
Server-based RAID storage is commonly available in configurations ranging from eight disks up to 48 disks, so it is ideal for small to midsize business and departmental networks. Because the system is not a “true” external storage appliance per se, the selection and quality of the components is left up to the manufacturer, vendor or user. The quality of the components plays an important role in the overall integrity of the system, reducing the multiple points of failure found in a server-based RAID solution. Prepackaged solutions offer better integrity and support, but still rely upon the quality of the integration and components. Common user tasks can be greatly enhanced through the use of third-party commercial software.
Single- and dual-controller external systems
In single-controller external RAID systems, there is no dependable cache recovery mechanism that protects data against system failure without the use of a dedicated battery backup module. However, external storage arrays with dual-active RAID controllers can provide reliable cache recovery through the use of mirrored caching. During operation, the dual-active controllers share the workload; however, if one controller fails, the second controller assumes full control. When using dual-active RAID configurations, data is written to the cache in both controllers before the write is acknowledged as complete. Then, if a controller fails, the second controller automatically assumes the workload from both the failed controller and itself, completing the write operation that was in process at the time of the failure.
Whether installing a PCI, external or server-based RAID, the most important thing to remember is that RAID is not a replacement for having a proper backup and archive plan to protect valuable data.
Luis Rodriguez is director of sales and marketing for Proavio USA, an A/V brand of Enhance Technology.