What is application-aware storage?
At the highest level, application-aware storage should recognize an application and be able to enhance its storage performance by improving the relationship between storage and applications -- often manipulating characteristics, such as data layouts, caching behaviors and quality of service (QoS) levels. While some storage platforms may be tailored to application-aware storage tasks, it is more common to see software tools imposed on top of existing disk storage resources.
However, a more precise definition is elusive because application-aware storage is not a single product or technology. This leaves a lot of room for marketing hype, multiple meanings and interpretation. "The definition lies in the eyes of the beholder," says Greg Schulz, senior analyst and founder of the StorageIO Group in Stillwater, Minn.
Application-aware storage can mean content awareness -- using APIs to tailor a storage system, such as an EMC Corp. Centera, for certain applications. However, the most common definition of application-aware storage today is strongly reminiscent of storage management -- using storage, storage systems or storage management tools to support policies, allocation, configuration and use -- based on application use or service requirements and profiles. "What this [application-aware storage] brings is better visibility to where the applications map to the storage resources," says John Merryman, principal consultant, data classification services at GlassHouse Technologies Inc. in Framingham, Mass.
In this context, application-aware storage presents a new "policy management" dimension to tiered storage where application-aware tools can use policies and business rules to automate the assignment of storage for specific applications. For example, messages and data generated by an email server may be mapped off to midperformance drives running RAID 5, while data produced by a backup application may be sent to storage provisioned at high-capacity SATA disks in a RAID 6 group. As another example, Oracle data could be sent directly to a high-performance Fibre Channel storage area network (SAN).
What are the benefits and drawbacks to application-aware storage?
Conceptually, application-aware storage is intended to improve storage performance and storage service delivery and automate some of the manual provisioning processes that have routinely demanded IT knowledge. The challenge is to establish a tangible relationship between the storage, the applications and the business policies. Storage professionals could then see those relationships and tailor them, allowing tools to make higher level decisions about storage provisioning. When properly deployed, a storage administrator doesn't need to worry about where an application is storing its data -- tools will make those decisions and distribute data accordingly to the proper storage platform, storage tier and LUN. Experts note that such capability will be increasingly important as storage resources proliferate and become more complex. "Institutional knowledge is deteriorating," Merryman says. "So having the technology to gain these insights is important."
There are numerous vendors today providing management software tools with "application-awareness" features. Software vendors include SANscreen from Onaro Inc., WysDM for Fileservers from WysDM Software Inc., StorageConsole from Aptare Inc., the Virtual Recovery Engine from Illuminator Inc., Veritas Storage Foundation from Symantec Corp., BalancePoint from Akorri Inc. and StorageGRID from Bycast Inc., among others. On the hardware side, experts point to EMC Corp.'s Centera CAS system, Hitachi Data Systems' Archivas platform and the Assureon archiving platform from Nexsan Technologies Inc. as examples of application-aware storage for archiving purposes.
The problem with application-aware storage is that practical products are still in the early stages of development and deployment. It's a technology that is really only suitable for large storage organizations, so it's still considered to be a niche feature. "Like any technology, the early adopters are paying the price to understand the value proposition," Merryman says. "It's probably not going to be the silver bullet to cure all tiering and [storage] efficiency woes."
Consequently, application-aware storage deployments can often be marred by unrealistic expectations of the technology. Tools still need to be deployed, configured, tailored and tweaked. Experts say it's important to evaluate and compare tools very carefully prior to deployment -- a pilot program can reveal many potential problems. "Define your requirements and lead with that," Merryman says. "Don't look at the vendor .pdfs or let that drive your requirements." Understand any interoperability problems, procedural changes, management overhead and scalability issues that present problems after installation.
How do you deploy application-aware storage?
Generally speaking, application-aware storage is a policy manager implemented as a software layer installed on top of storage, and on top of storage management utilities. The policy manager extracts the established technical rules and business policies necessary to make rudimentary decisions. "A policy manager is no good unless it's got good, viable rules and information to act upon those rules," Schulz says. "This allows the policy manager [application-aware storage software] to act upon those rules."
Once installed and configured, the application-aware storage software should operate with virtually any storage platform in the enterprise. Experts emphasize that any storage system has the potential to be "application aware." Still, it's important to verify interoperability with your current storage systems and existing storage management tools before fully committing to an application-aware storage product.
Also, you can employ purpose-built storage hardware to form the foundation of an application-aware storage infrastructure. For example, the Sun 5800 requires WebDAV or an API to access the data objects that are stored on its disks, thus mating the accessing application to the storage system. Similarly, EMC's Centera relies on the use of APIs that tie directly to the applications.- Some experts refer to this as "application-affinity" or "application-affiliated" storage. "You just can't mount a volume with NFS, and sit reading and writing the files," Schulz says.
Remember that application-aware storage is typically intended to merge technical and business practices, so implementation cannot be an IT-only process. Successful deployment must involve participation from throughout the organization to understand the importance of applications, the data that applications produce and the optimum rules/policies for storing that data -- a process similar to data classification.
Metadata also has a role to play in application-aware storage, using details about various data types to feed the policy managers that make decisions about where and how that data is stored. However, it's important to note that this technology relies on very different details than the metadata used in traditional indexing or e-discovery searches. A CMDB (configuration management database) provides metadata details about the systems, applications and their configuration, rather than details about the application's data. "You're dealing with infrastructure resource management (IRM) -- managing the resources to deliver IT services," Schulz says.
What is the future of application-aware storage?
Application-aware storage is certainly not a new idea. It has been mentioned as far back as 2003 and earlier, but has only recently started to receive attention as a unique and viable storage technology. Experts suggest that this recent acceptance can be traced to several factors, including an almost universal use of SAN technology and a fundamental shift toward server and storage virtualization to improve infrastructure efficiency, reducing the tendency to overbuy storage to solve availability and capacity problems.
The most notable issue is a broad lack of deployment, so experts are still hoping to see more active rollouts of application-aware storage tools. Beyond that, the goal is greater automation, allowing the software to act on wider range of policies and transparently make more autonomous decisions about storage in the enterprise.
This was first published in August 2007