Broadcom The Challenge of Protecting Data with RAID
The challenge of protecting data with RAID
It’s an imperative for organizations of all sizes: data storage must be fast for speedy access and safe, protected by redundancy. To safeguard data, most small and medium businesses (SMBs) and enterprises (SMEs) use some form of RAID (Redundant Array of Independent Disks) data protection across server storage assets. RAID mirroring, striping and fault-tolerant parity are combined in various ways in order to tune the performance, data redundancy, availability, cost and storage capacity to the criticality of the data. RAID parity is a key technique used to protect data, reconstructing a RAID volume if one or more drives within an array fail. In general, the higher the RAID parity, the higher the computational load.
Parity computation is performed either by software that runs on the host system’s CPU, requiring precious processing power, or a discrete storage controller card that spares the host the additional compute load. The advantage of software RAID is lower cost since, in most cases, no additional hardware is required. But using software for advanced RAID data protection schemes for enterprise workloads, such as RAID 5 and RAID 6, can impose a heavy processing load for parity computation, sapping the host of significant processing power and even degrading the performance of other applications, reducing return on investment (ROI).
RAID solutions that rely on attaching drives to integrated Serial Attached SCSI (SAS) and Serial ATA (SATA) also threaten to reduce ROI. Many server motherboards support only four or eight hard disk drives (HDDs), a natural configuration restriction on storage growth. Businesses adding more storage must manage multiple volumes across different controllers, complicating management, draining IT time and risking incompatibilities among storage solutions.
Compared to software RAID, hardware RAID based on a dedicated storage controller delivers far superior storage performance. The ROI with hardware RAID is also much higher since the host compute resources are free to conduct other line-of-business tasks.
The MegaRAID® family of storage controllers is designed to meet critical requirements for RAID data protection and server storage ROI. The family:
- Enables the server to boot directly from the RAID array, eliminating the need for specialized boot drives and protecting information – such as normal operating system and application operations – not kept in main storage
- Allows IT managers to standardize on a single platform-agnostic controller that supports large, highly scalable drive pools to simplify storage expansion and management
- Supports a wide range of RAID levels including RAID 6, RAID 50 and RAID 60 to accommodate growing business needs for data protection
- Supports caching acceleration options for higher levels of RAID data protection to optimize performance. MegaRAID high-performance, non-volatile DRAM write caching is designed to significantly reduce disk bottlenecks during high-throughput burst activity.
- Supports the use of solid state drives (SSDs) to cache and accelerate frequently accessed data, combining the benefits of high-performance SSDs and high-capacity HDDs.