The debate over storage intelligence has been raging for years. Should the storage brains reside on the server with the operating and file system, or should it live on storage subsystems with the data and disks? In the midst of this battle, a third option is steadily gaining ground. Because storage architectures are now tightly aligned with networks, why not put the intelligence in the fabric? After all, LAN equipment delivers software smarts for activities such as routing, QoS and security, so why not arm storage area networks (SANs) with brains of their own?
According to a current research report by the Enterprise Strategy Group (ESG), Milford, MA, the time for intelligence in the fabric is now. ESG recently surveyed 210 IT professionals to gauge their thoughts on intelligent fabrics. An intelligent network was defined as "a storage network with a resident switch or appliance which performs storage services (e.g., replication, virtualization, continuous copy) or in which storage software, applications reside." Based on the results of our survey, 21% of the companies have already implemented intelligent storage fabrics and the companies that haven't express a strong interest in doing so.
Who's getting smart?
What types of companies have taken the plunge? The storage elite. Early adopters tend to be large companies with oodles of storage, big budgets and complex SAN architectures. These kinds of firms have the highest frustration levels, but they can achieve the greatest benefits by moving intelligence onto the storage network.
According to the survey, the top reasons to implement an intelligent fabric are to simplify storage resource management (76%); reduce hardware spending (49%); and improve the ability to provision storage capacity (47%).
In spite of the relative youth of intelligent storage fabrics, this growth trend shouldn't surprise anyone. While large companies implement intelligent SANs, they're also consolidating data centers, flattening network architectures and aggregating security functionality. The 1990s left us with a hangover of too much equipment and not enough time or manpower to manage it. Intelligent fabric technology is yet another appealing way to rationalize the IT infrastructure to improve service and lower costs.
Have intelligent storage networks led to results? The evidence suggests the answer is yes. Regarding the three reasons for implementing an intelligent fabric stated earlier, companies had either met or were close to meeting their objectives. There also seems to be some surprising benefits such as a newfound improvement on deploying storage services and a reduction of interoperability issues.
In addition, 79% believe that intelligent fabrics enable them to reduce their storage hardware spending to some degree, with similar results for storage software spending and administration costs. It's also worth noting that when asked to compare initial expectations with actual outcome, early adopters say they achieved better than expected savings.
A few major adjustments
These results are certainly encouraging, but remember that implementing intelligent storage fabrics is an architectural change, and not for the faint of heart. Storage professionals need to research solutions and understand what type of impact they will have on infrastructure and operations. Performance and availability requirements must be considered as well as which vendor products can be eliminated. This can turn into political football.
And there's more. Budgets must be estimated and approved. Vendor products must be researched and tested. Consolidation projects must be managed effectively and on a tight schedule. Finally, storage personnel must be trained on new products and have processes for reacting to inevitable problems.
A project with this much detail is bound to have its set of issues. When early adopters were asked to name their top challenges, the top five listed were: high acquisition costs; changing existing IT processes to match new capabilities; interoperability with existing infrastructure; deployment difficulty; and high operating costs. These types of problems aren't unusual for a new technology initiative. The journey may be difficult, but it's worth the effort.
As f0r which vendors stand to benefit from the intelligent storage trend, it appears that the rich will get richer. Users don't want to piece together components. Rather, they want to purchase integrated products from vendors that offer strong support. They also voiced a strong desire for scalability, with future plans to have multiple storage services on each intelligent system. Potential vendors included Brocade, Cisco, EMC, HP, IBM, McData and Veritas. Startups may have some appeal with later adopters who are willing to look at a broad array of solutions.
Preparation is key
The results of the survey show that fabric intelligence is no longer a vision or technology argument--it's a trend that's already producing results for leading companies. Yet fabric intelligence is still a nascent movement, as only 21% of ESG's survey respondents have deployed solutions. As for the other 79%, it's important that they prepare for intelligence in the fabric by:
Not purchasing host- and storage-based products. These products may offer attractive functionality and maturity, but chances are strong that today's host and storage software will require a migration to the network within the next three to five years. Furthermore, storage software R&D will focus on network intelligence, so the latest and greatest features will arrive on the fabric first. Postpone software purchases until you're comfortable with network-based solutions. Short-term pain will be compensated by long-term flexibility and savings.
Planning to start slow and grow. Resist the temptation to go gaga over storage intelligence and take a much more pragmatic approach. Pick the one application that will serve the business or greatly reduce costs as a starting point. This will keep the architectural, training and deployment challenges reasonable and make the next application project flow more smoothly.
Investigating consolidation options. While easing into network intelligence, look for creative ways to simplify the architecture and eliminate vendors down the line. For example, can a single network package replace multiple replication packages? What would the ramifications be? Explore these alternatives to take full advantage of potential gains.
Vendors should plan on porting their software to leading intelligent storage switches, blade servers and appliances. During this heavy development period, storage firms must also:
Educate their customers. Many firms remain confused about intelligent fabric technology. Vendors need to build trust by schooling customers through presentations, training material, road shows and professional services.
Outline a migration path for legacy systems. Customers will want to know how vendors will hold their hands though the migration process. Smart vendors will make this a win-win by minimizing customer disruption while creating new product and services revenue opportunities.
Define new service options. One overwhelming user concern is service. What happens when multiple applications are running on a storage switch? How will they upgrade these systems? How will they test new application versions? Will vendors cooperate? These are legitimate questions that may be foreign to vendors that are used to only servicing their own equipment. Users need answers here, so vendors have some work to do.
The results of ESG's intelligent fabric survey paint a clear picture: The storage network will get smarter over the next few years, and users are willing to deal with architectural and process shortcomings to achieve their goals. This trend will force the storage industry to take a network-centric approach to their R&D, customer service and sales efforts. Strap yourself in, it's bound to be an interesting couple of years!