Building storage networks that are "application-aware" enables IT organizations to enforce service levels and availability policies based on the business value of data and applications.
By Rick Walsworth
In these challenging economic times, IT professionals are often faced with the mandate to protect their company's expanding data infrastructure while trying to maintain fiscal responsibility—requirements that are typically mutually exclusive. Intelligent data centers built with storage infrastructures that are application-aware enable storage administrators to deliver non-stop availability to the applications that require it; in the event of a disaster, they would also cost-effectively protect all of the data within the network based on the business value of that data.
In an effort to consolidate server and storage server infrastructure to reduce costs, many storage professionals have deployed storage area networks (SANs) that provide access to common storage resources, increasing resource utilization and leveraging the SAN for high-end applications. Although SANs provide consolidation of physical resources, these layer 2 infrastructures have no intelligence about the service-level requirements of applications accessing the data.
Application-aware storage networks provide the ability to establish unique service levels and availability policies for the applications sharing the storage infrastructure. Once these policies are set, applications can be provisioned and services can be rapidly deployed in a uniform, consistent manner. Using storage quality of service (QoS) to enforce these policies within the network ensures that the service levels are maintained throughout the infrastructure in the most cost-effective manner.
In order to implement cost-effective business continuance strategies in heterogeneous storage environments, it is important to define your strategy to dynamically enforce ser-vice levels and availability policies that are based on the business value of the application. By building storage networks that are "application-aware," you can realize benefits by matching the disaster-recovery policies to the value that each application brings to the business, as shown in Figure 1.
Figure 1: Application-aware storage networks enable IT administrators to match disaster-recovery policies to the business value of applications.
Because not all applications are created equal, they fall into different categories with different storage and service-level requirements. Workgroup applications may generate large amounts of information that have specific performance requirements. Externally networked applications such as supply chain management extend not only throughout the enterprise but also to customers and suppliers. Data-base applications have specific requirements for latency and throughput. Each of these applications has unique characteristics and places different demands upon the infrastructure for scalability, availabili-ty, and performance.
Regardless of their scope, applications that directly impact revenue generation are mission-critical and must be assigned the highest priority levels of protection; downtime for these applications is not an option. While other applications may not directly contribute to a company's revenue stream, downtime would result in lost productivity or impact customer service; these applications require adequate protection, but perhaps at different investment levels.
In business continuance planning scenarios, data-center architects commonly prioritize the applications that will be protected, putting those applications on their large storage subsystems and the other applications on whatever is left over. This task is more complicated in a consolidated data center, where prioritizing data protection becomes a cost-prohibitive, time-consuming, unreliable, and often political process.
To prioritize disaster recovery across the enterprise and provide the highest levels of protection for critical applications, the SAN must be able to classify storage resources and allocate them as pools of tiered storage. To do so, the storage switching fabric must be able to differentiate the data flows and associate service levels with applications.
Since current SAN switching technology processes data at layer 2 of the protocol stack, it lacks intelligence about the application requirements. For this reason, storage networks are following the same migration path to network-based intelligence as data networking products did more than a decade ago.
Data networks evolved from a layer 2 Ethernet LAN to layer 3 and eventually to layers 4-7. As the networking equipment moved up the protocol stack, network-based intelligence enabled services such as dynamic bandwidth assignment, QoS, security, and application-based load balancing to be enabled from the network. As storage networks become application-aware, the intelligent fabric has the ability to examine the contents of data flows and enforce assigned service levels directly in the data path without performance degradation. This capability enables administrators to develop and execute a business continuance strategy that matches the value of the applications to the business.
With these capabilities, administrators define appropriate storage service policies and what type of storage will be used to connect to the application servers for primary, secondary (mirrors), and tertiary (snapshot) data volumes. The application-aware infrastructure then acts on these definitions by dynamically allocating the appropriate storage resources, examining the data flows, and applying the appropriate levels of protection based on an application's service levels. This infrastructure enables administrators to rapidly roll out disaster-recovery and business continuance implementations cost-effectively, particularly when using less-expensive disk arrays for mirrors and snapshots (see Figure 2).
Figure 2: With an application-aware storage infrastructure, administrators can define policies that dictate what type of storage will be used for primary secondary (mirrors) and tertiary (snapshot) data volumes.
An intelligent data center built on an application-aware switching infrastructure can provide the following benefits:
Lower cost of ownership—By using cost-effective storage arrays for mirrors, replicated volumes, and snapshots, capital costs are significantly reduced. By providing a policy-based model for deploying these services, complexity is removed and management costs are reduced.
Increased data protection—By assigning service levels to applications, critical data is backed up and managed so that it can be quickly and completely restored in case of a disaster.
Efficient resource utilization—With storage QoS, storage network resources are dynamically allocated and more efficiently utilized, freeing up resources for additional data protection.
Higher consolidation levels—With network-based intelligence and data protection, an application-aware infrastructure provides the benefits of a consolidated storage network, but the QoS delivers the assurance that application performance or availability will not be impacted.
Application-aware storage networking helps organizations cost-effectively leverage the full potential of their SANs, providing ways to leverage their entire storage infrastructure to roll out business continuance strategies that fit the economic requirements of the business.
Rick Walsworth is director of product marketing at Maranti Networks (www.marantinetworks.com) in San Jose, CA.