SATA for business critical applications: the mirroring advantage; Affordable, mirrored SATA drives overcome FC RAID-5 obstacles.

As the volume of business information explodes, companies are struggling to find cost-effective solutions for storing their valuable data. The combination of exponential data growth, tight budgets and the complexity of data management has, in the past, required companies to deploy RAID solutions that offer protection with a price--in terms of both cost and performance. But now there are data protection solutions that deliver affordable alternatives (without performance degradation) using mirrored SATA drives to ensure data integrity.

For utmost data security, in the past, companies typically chose to deploy Fibre Channel (FC) storage implementations with RAID-5 protection. While this scenario allows the ability to rebuild any one failed drive using parity information, it comes with major costs in two areas: direct hardware costs and performance costs. But now, with the advent of lower cost Serial ATA (SATA) disk drives and the increased amount of intelligence and enhanced data services available in SATA storage solutions, IT administrators have a cost-effective alternative to protecting their data.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

Let's look at a typical RAID enclosure that utilizes RAID-5: four of the six Fibre Channel drives are data drives, there is a parity drive (although parity is rotated) and a spare drive to round out the set. Now imagine one of the data drives fails. The storage system begins the rebuild of the failed drive to the spare and, as information is restored, performance of the entire system grinds downward. According to Murphy's Law, this rebuild will probably take place at an inopportune time, when the data on the drives is needed for critical applications. And since there is usually no real control over when the array begins rebuilding, you have no recourse except to grind your teeth, watch the rebuild and get ready for the phone calls. That is, until now.

SATA drives have now matured to the point where they allow companies to deploy cost-effective mirroring solutions as an alternative to costly FC RAID-5 implementations. Using low-cost SATA drives for primary and secondary storage, primary data sources can be mirrored in a drive-for-drive implementation to create an identical copy of each drive, and this solution can deliver numerous advantages to companies choosing this approach.

The Challenge: SATA vs. FC

Let's examine a typical data-protection strategy that would deliver the same level of protection as the Fibre Channel example above, but using SATA drives. Three SATA data drives are used for primary information, with no need for parity or spare drives. Another three identical SATA drives would mirror data from primary sources, in a one-to-one relationship.

Now imagine a failure of one of the primary SATA drives. In this scenario, administrators simply use data on the mirrored secondary drive to continue immediate operation with no rebuilding needed. There is no performance "hit" to the system because rebuilding does not need to occur. Also, overall hardware costs are much lower using SATA drives. And with mirroring, you're in charge of choosing when to replace the failed drive during off-peak system usage.

At this juncture, and with these common scenarios, SATA provides cost and performance benefits. Now, let's weigh the challenge of two drive failures on each system. With SATA mirroring, your system is still online and operational (unless you manage to have an identical pair of mirrored drives fail, which is a very unlikely occurrence). With RAID-5 and two drive failures, you're completely off line and unable to restore from parity. With a RAID-5 Fibre Channel implementation, you will be restoring from the latest backup and watching productivity (and profitability) go out the window.

As you can see, using a mirrored SATA approach to data protection brings a number of immediate benefits for storage professionals, including:

* Gaining a cost-effective advantage at deployment, with lower initial storage hardware investment

* Having backup information immediately accessible from the mirrored SATA drive, without the need for lengthy rebuilds

* Avoiding the performance degradation that occurs during rebuild operations

* Securing the ability to choose when to effectively replace damaged drives, minimizing impact on business activities

As the price of SATA disk drives continues to drop, and as companies continue to increase the features and functionality of the integrated SATA solutions (including the addition of enhanced enterprise-level features), SATA shows the potential of becoming the storage architecture of choice for more and more businesses that are concerned with deploying resilient, yet cost-effective storage solutions to meet corporate objectives. Now, let's explore the attributes of SATA storage solutions that can return value and allow for future infrastructure growth.

SATA Characteristics That Make a Difference

As you assemble the necessary features and functionalities needed to create mirrored SATA solutions for corporate environments, the following characteristics of enterprise-class storage solutions that have returned value and cost savings to large and small data centers will prove useful in selecting SATA solutions for larger roles in your data center. When reviewing SATA solutions for your particular implementation, having these characteristics can mean the difference between a system that provides a mere stopgap solution and one that provides a flexible and robust growth path for continuous return on investment.

Uncompromised Levels of Service: In today's mission-critical computing environments, service level agreements (SLAs) commit IT organizations to measurable service metrics such as application performance, end-user response time and system availability. Since storage can directly impact the performance, availability and functionality of most applications, storage plays a vital role in an IT organization's ability to deliver IT services.

Data centers, large and small, require solutions that provide uncompromised levels of service, backed by quality support and a migration path to new technology that protects current investments. They also require storage solutions with robust architectural designs that withstand potential failures without impacting data availability. Top data center reliability, availability and serviceability (RAS) features for SATA solutions should include items such as no single point of failure, redundant I/O paths to SATA drives, high-performance RAID for protection against disk failures, proactive error detection and response (memory scrub and cache vaulting, for example), the capacity for nondisruptive operating environment upgrades, enclosure environmental monitoring, and a modular design enabling nondisruptive component replacement.

Additional Data Services for Data Asset Protection: While the RAS features and redundancy techniques above are vital and necessary components for a storage solution, they only protect against hardware component failures, and for this reason, are insufficient. Beyond hardware failures, disruptions to data access can also come from human errors, data corruption or natural disasters. And because companies increasingly rely on information access as the lifeblood of their business, disruptions to data access puts the entire enterprise at risk. This risk can be quantified in a number of ways, including lost revenue and lost business opportunities.

Storage systems, sometimes seen as a container of corporate data assets, have emerged as a critical resource that has evolved beyond a traditional capacity model and should be viewed as more than just a receptacle. With today's advancements in network technologies, storage is also a platform for data services to manage and mitigate risk to data access. A tightly integrated combination of storage hardware, embedded intelligence and data services are the means to practical risk management and data asset protection; a storage solution that supports business continuance, backup/restore and capacity management helps to ensure that a company's core business activities continue to function with minimal interruption, regardless of catastrophic events, malicious attacks, upgrades or common everyday mistakes. In addition to mirroring, companies looking to suitably protect their information, maintain data availability and reduce risk management should include:

* Snapshot services to provide the ability to recover from logical errors (such as corrupted data or deleted files) by preserving data at multiple points in time.

* Virtualization of physical disks to enable provisioning of storage on demand, thus increasing system availability. By enabling the consolidation of storage, management complexity and costs are reduced while overall capacity usage is enhanced.

* Capacity management services to help avoid costly downtime due to unforeseen changes in storage capacity requirements, such as file system growth. By monitoring and automatically expanding storage space from available capacity, downtime due to physical disk upgrades, rebuilding and reformatting of operating systems is eliminated.

Common Management Tools: Storage management software serves as the foundation for networked storage, enabling IT organizations to tie together all their storage resources, to monitor them and to control them. To maximize return on investment (ROI), it is important that the configuration, analysis, and control be provided by a central, easy-to-use graphical user interface (GUI) that can be securely and remotely accessed anywhere.

Investment Protection: As companies invest in new storage systems and technologies to meet the needs of their growing IT infrastructure, it is essential that they deploy storage solutions that will evolve with their business needs. This is important, since deploying a storage solution developed with long-term product continuity in mind will return the highest level of investment protection. Companies can leverage storage solutions to maximize ROI, protecting their initial investments in hardware, software and training.

Key areas that such a storage solution should address include:

* Modular interfaces which offer multiple connectivity choices (e.g. FC or iSCSI and the ability to migrate between protocols with data intact).

* Modular architectures, which enable performance upgrades to newer and higher performance controllers. A seamless upgrade with data intact eliminates the downtime and risk of data migration.

* Hardware and software compatibility across platform generations, which eliminate the burden of retraining IT staff.

* Compatibility and data movement across hardware generations to enable older systems to be utilized at secondary disaster recovery sites as new systems are used for production.

* Common reusable components (such as storage enclosures, disk drives, and power supplies) to preserve a majority of hardware investments across storage platforms.

Modular Storage: Choosing a modular architecture offers a "build as you grow" approach, such that as more capacity is needed, additional drive shelves can be added non-disruptively. This approach lowers initial acquisition costs for cost-conscious environments such as SMBs (small to mid-sized businesses), and can also allow demanding environments that require greater I/Os, throughput and capacity, to easily scale as requirements grow.

The modular storage concept also applies to the data services components--the intelligence--of an intelligent storage system. The ability to license data services components on a usage and capacity basis makes storage system acquisition costs more affordable while providing a path for future growth requirements.

Native Command Queuing Support: NCQ is a widely anticipated feature introduced in the Serial ATA II extension to the Serial ATA 1.0 specification. This command protocol allows multiple commands to be outstanding within a drive at the same time and support for NCQ allows drives to dynamically reorder command execution and optimize the workload while minimizing mechanical latencies. It follows that intelligently managing the sequence of mechanical motions not only improves performance, but can also reduce mechanical wear and enhance endurance. A storage solution that supports and leverages the NCQ feature in SATA drives delivers higher overall system performance, as well as greater reliability in the storage infrastructure.

Commitment to Industry Standards: The Storage Network Industry Association (SNIA) has developed and released the first set of specifications to provide common models for devices and a common protocol for managing storage networking devices. The SMI-S (v1.0) industry standard defines a model for discovering and managing arrays, switches and HBAs. SMI-S also makes it possible to eliminate the need for vendor-specific APIs and to do away with the myriad of management protocols used today to manage storage environments.

The benefits to end users from the SNIA initiative are significant. End users benefit through an increased flexibility to choose both hardware and software vendors. Vendor lock-in is eliminated because SMI-S makes it easier to implement heterogeneous solutions. Implementation times are reduced as new storage devices seamlessly integrate into the management interface, and end users will be able to realize both management and operational cost savings by managing heterogeneous solutions as though they were from a single vendor. Given all these benefits, support for SMI-S should therefore be important to end users as they make purchasing decisions regarding storage and management solutions.

SATA: An Affordable Vehicle for Intelligent Primary Storage

SATA technology has matured to the point where it is an ideal solution for the workload, capacity and cost constraints of businesses wishing to deploy a robust, intelligent and affordable storage infrastructure for data protection. When properly implemented, SATA technology can be used as the foundation of a cost-effective solution that can compete with Fibre Channel RAID 5 solutions, and which brings greater value to the data center.

In developing a new generation of intelligent data storage solutions to address and serve these primary storage needs, companies should be committed to incorporating intelligent, enterprise-level functionality into all of their solutions (including SATA) making advanced data services features available to organizations that can realize the most benefit from cost-effective solutions.

With the exponentially increasing volume of business information, the use of intelligently mirrored SATA storage solutions can bring a cost-effective answer to address the enterprise-level needs of a wide range of businesses. With intelligent SATA solutions, businesses can address the exponential data growth that plagues data centers around the world. Using intelligent SATA, businesses can deploy storage networks with enterprise-level functionality at a reduced price. And with intelligent SATA, businesses can protect information with smart, tiered approaches to address data protection, and gain performance benefits with a solution that embraces a strategy of long-term product continuity combined with the highest levels of investment protection and workflow efficiencies.

Jonathan Kong is executive vice president, CTO, and cofounder of iQstor Networks (Newbury Park, CA)

www.iqstor.com