This SAN configuration tip is the third in a three-part series about the role of the storage area network in midmarket organizations. In the first article, the various types of SAN solutions were explored. In the second part of this series, the author dives deeper into the individual capabilities of SAN technologies.
After you've determined your storage requirements and evaluated the vendor landscape, it's time to move on to the next phases of your SAN investment: acquisition, installation, configuration and ongoing management.
The decisions made during the SAN selection process should accurately determine how easily the solution will integrate into your environment. Management tools should help facilitate rapid setup, configuration and initial deployment, and even some of the ongoing, day-to-day operations.
But one of the questions that frequently comes up during the initial SAN configuration process is choosing the appropriate RAID level necessary to meet your various needs. While wizards and other setup tools can help automate this task via a series of questions, templates or examples, the following is a quick and general synopsis of the RAID levels required for different purposes:
RAID 0: Boosts performance but does not provide data protection, and loss of one disk results in disruption across the entire RAID group. At this level, data protection is traded for performance and cost. Use for applications where data can be restored quickly from some other medium, including disk-based backup with deduplication or tape.
RAID 1: Provides good performance and availability by duplicating data on two or more disk drives. However, this duplication also reduces capacity. Use for read- and write-intensive application data, including database tables, indices or journals where downtime is not an option.
RAID 4: Good performance and data protection. At this level, performance, capacity and data protection is balanced. Use for general-purpose file sharing and other common applications ranging from email to video, home directories, etc.
RAID 5: Similar to RAID 4, RAID 5 balances performance, capacity and data protection. It's well suited for concurrent reads and writes when combined with write cache. Use for general-purpose file sharing and other common applications ranging from email to video and home directories.
RAID 6: For less frequently accessed data, this level provides reasonable performance similar to RAID 5 with an additional parity drive for extra data protection. Using larger-capacity disk drives in RAID groups with more disk drives can help you achieve a balance of good read performance, reduced parity (or data protection overhead to usable capacity), cost and a lower-impact large disk drive rebuild time. Good for archives, online reference, home directories or other applications where low cost of storage is needed with good data protection and availability.
In addition to initial setup and SAN configuration, other early-stage tasks include stabilizing snapshot schedules and integrating applications such as Exchange, SharePoint, Oracle, VMware or Hyper-V. Ongoing tasks will include performance monitoring, diagnostic testing, storage provisioning, allocating space on new servers and moving data during upgrades. Integrating data snapshots with backup processes for disaster recovery and business continuity data protection efforts will also be part of the routine.
Basic skill sets required to perform many of these tasks include familiarization with servers, operating systems, storage and networking. Some SAN providers also advertise their products as being able to be managed by a CEO or CIO. While this may not be a key requirement for you, don't discount ease of use or automation and scripting capabilities, all of which can reduce some management burdens.
The importance of SANs and server virtualization
Server virtualization is becoming more prominent in midmarket environments. While there is no hard, written rule stating a SAN is necessary for virtualization, some form of shared storage is required if you plan on having two or more physical servers.
If you do decide to use a SAN to meet the shared storage needs, keep hypervisor (e.g., server virtualization software) integration and end-to-end management capabilities in mind when selecting your product. Make sure that the solution works with and is integrated and supported by one of the hypervisor solutions on the market, such as Microsoft (Hyper-V), VMware Inc. or Citrix Systems Inc. End-to-end management capabilities are also important because they remove some of the day-to-day management complexities associated with a growing storage environment by enabling overall management capabilities of virtual servers, storage networking equipment, storage devices and associated software.
So what's the bottom line? A SAN should work for you, and not the other way around. Balance the ongoing workloads by taking advantage of management tools, especially if virtualization is in your plans.
Greg Schulz is founder of The Server and StorageIO Group, an independent IT Industry advisory and consultancy. He has authored the books The Green and Virtual Data Center (CRC) and Resilient Storage Networks. Find him on Twitter @storageio.