It's easy to understand why content-addressed storage (CAS) systems have attracted such close attention. Most companies must manage a huge amount of fixed content and that content is growing at an alarming rate each year. CAS provides an economical storage platform, which can offer far faster service than traditional tape, and can help organizations of all sizes to meet their regulatory obligations. The big question then is "where do we go next?"
There is little doubt that CAS storage capacities and object counts will continue to grow over time. Yet even as systems escalate in capacity, their value will benefit from falling storage prices -- especially with midtier SATA and SAS hard disks. Analysts also see CAS technologies specializing in more specific niches.
"I expect to see this type of technology becoming almost a standard feature of VTLs [virtual tape libraries] because it helps drive the cost of disk closer to the cost of tape," says Jim Damoulakis, chief technical officer at GlassHouse Technologies Inc. "It's also likely to be applied more extensively in backup realms. The idea of reducing the amount of redundant data that gets backed up is very compelling."
Analysts like Greg Schulz, founder and senior analyst, StorageIO, expect that CAS products will extend their reach into high-end and low-end markets, while becoming easier to use, and offering better integration with operating systems and application software. "I think there can be tighter integration with operating systems," he says. "Maybe some of that object- or CAS-based functionality moves up into the operating system or into the virtual operating systems."
The future evolution of CAS will also depend on the development of newer, more powerful applications, such as search and data management tools. "If the archive storage system grows to millions or even billions of files, then searching for the data must be easy," says Tony Asaro, senior analyst at the Enterprise Strategy Group. "If there is a discovery process for litigation or compliance purposes, then it is business critical to get any and all data that is evidentiary. Since there may be many files created by different applications, a common search engine can be extremely useful to provide fast access." Today, such tools are typically vendor specific, but look for a richer suite of software that is interoperable across a wider range of CAS products. This will probably become a reality once CAS vendors embrace a standard interface method.
The move to standardization
In order for CAS software options to expand, CAS vendors must adopt a standard interface that can allow a software product, such as a metadata search tool to operate across numerous platforms. "There should be standards that allow customers to move from one CAS system to another, regardless of product or vendor, and retain the policies, WORM [write one, read many] and retention periods," Asaro says. "Right now, there is massive vendor lock in."
The problem today is that every vendor uses an application program interface (API) to create and store metadata, so applications must be coded to suit each API. This adds complexity and expense to the software development process, and causes incompatibilities between features. However, that may change if the Storage Networking Industry Association (SNIA) has its way.
"The Fixed Content Aware Storage Technical Working Group is defining one API for all storage vendors to use for this kind of storage," says Mark Carlson, chair of the SNIA Technical Council. In September 2005, industry leaders like IBM, EMC Corp., Sun Microsystems Inc., Hitachi Data Systems Inc. and others submitted a proposal known as the eXtensible Access Method, dubbed X-Access Method or XAM. The objective is to ratify a single interface -- one API -- that supports CAS-type systems without regard for the storage technology, the vendor or the location of that storage in the enterprise infrastructure.
According to Carlson, the workgroup is meeting regularly and he expects a standard sometime in 2006, with vendor implementations by 2007. "The real key here is to get the operating system vendors to include this API as part of the operating system so that you don't have customers installing software before they can use their [CAS] storage," he says. SNIA is considering a software developer's kit to help software developers with CAS application development.
Final thoughts and advice
Analysts generally agree that CAS itself has little inherent value unless it solves a real-world problem, so it's particularly important for you to understand the business needs and implications of a CAS system. "At this point, the reason that customers are buying these solutions is not all that clear," Asaro explains. "On the one hand, customers are embracing WORM for compliance and litigation protection. However, others are just buying CAS solutions as a low-cost archive storage system." Knowing why you need a CAS system is always a vital question.
Don't hesitate to spend some quality time evaluating CAS products prior to a purchase, and ensure that any prospective purchase offers the scalability, integration and management features that meet your needs.
Any consideration of CAS should also include a serious evaluation of long-term storage realities. "Do you intend to have data stored on a CAS solution for 10, 20, 30 years or more?" Asaro says. "How big will your CAS system get? Is it practical to keep data that hasn't been accessed in years on spinning disks?" Addressing such issues before deployment can have a tremendous impact on long-term product selection as well as data management policies.
Go to the first part of this article: Introduction
Or skip to the section of interest:
- Content-addressed storage: An overview
- Content-addressed storage: Strengths and weaknesses
- Content-addressed storage: The vendors
- Content-addressed storage: User perspectives
- Content-addressed storage: Future directions