2003 | 2002 | 2001 | 2000 | 1999 | |
Storage Software | $6621 | $5730 | $6157 | $6113 | $4640 |
Storage Services | $23,360 | $21,171 | $20,552 | $19,501 | $17,250 |
Note - In addition to the "lumpiness" of the software revenues, it is also critical to note the steady increase in storage services revenue. These figures indicate strong growth in the use of professional services staff to install, configure, and manage storage hardware and software.
In the long run, software revenue figures are augmented by the eventual maturation of storage management products, a process that is still, at least in part, dependant on consolidation in the storage software market itself.
BMC Software's exit from the SRM market (and subsequent sale of Patrol Storage Manager to EMC), StorageNetworks' flameout, and EMC's purchase of Legato, Documentum, and Astrum increase the chances that products on the market eventually meet customer requirements.
These developments also indicate that additional consolidation is still on the horizon. As is typical of the churn around the chasm, product adoption rates and product innovation rates are "lumpy," showing spikes and valleys until the cycle rounds off with the early and late majority adopters.
A primary factor of the recovery of the enterprise storage software market is the demand for SRM software. SRM software is a fundamental component of any storage management initiative and eventually the one tool most capable of maximizing storage investments.
Storage Resource Management
Storage consumers have undoubtedly waited a long time for SRM tools to make their lives easier. Until SRM products mature completely, which depends in part on the ongoing development of standards for storage hardware products, storage managers and storage decision-makers remain dependant on a combination of home-grown tools and reports for capacity planning and provisioning. The tools available on the market today play only a small role in the day-to-day tasks required for the management of enterprise storage. Many customers who have already purchased and implemented SRM solutions are still reliant on traditional spreadsheet solutions for making critical decisions regarding the purchase, allocation, and de-allocation of storage. For SRM software vendors, the major competition is not other vendors, but their customer's own spreadsheets and processes.
SRM tools have high price tags, are notoriously difficult to implement, and in many cases, they have not delivered the critical reporting functionality needed to replace home-grown solutions at the heart of IT storage departments. Integrating SRM solutions with these tools presents a significant challenge to customers who cannot deploy labor-intensive software agents on a wider scale until the replacement solutions match basic functional reporting requirements. After these requirements are met, customers still face scalability issues with products designed to gather enterprise storage data from each of possibly thousands of hosts. Consulting and professional services in the SRM space are currently poised to see increased revenue because many SRM installations either fail or they require significant customization to show value.
In the not-so-distant future, the enterprise storage software market will bear little resemblance to the entity we are familiar with today. Marked by intense competition, rapid consolidation, and significant advances in technology, the enterprise storage software market will become an oligopoly, with a handful of major players commanding the majority of the revenue. Over time, smaller companies, such as Boston-based Onaro and Silicon Valley-based MonoSphere, that focus on what are now seen as niche markets (for instance, predictive change management and virtualization) will gain momentum and market share by supplying unique features and functionality that lessen the burden of storage management.
Currently the SRM market is, with a few exceptions, marked by a handful of immature products that show significant promise.
Reliance on storage management tool suites, when they finally mature, will simplify every aspect of the storage yield, purchase, and deployment cycles. Capacity planning will finally move away from spreadsheets and utilization, at the disk level and application level (the allocation and utilization efficiencies) will be tracked in real-time through a GUI. Performance, service level management, workflow, and information lifecycle management functionality will also be built into future storage management software packages. Until that day, however, many customers are in a wait-and-see position with regard to installing the current releases of SRM solutions. Consequently, they must rely on de facto standards, homogeneous environments, and home-grown solutions to make their storage management and operational lives simpler.
Most capacity planning and disk procurement processes are based on a combination of historical trending, usage models, and guesswork. These inefficiencies, coupled with the immaturity of storage management software, irrational exuberance, and Y2K-related spending, are all responsible for much of the overcapacity and build out of disk capacity in the years previous to the most recent recession.
It is important to put the current state of affairs into perspective, however. If you concede that Fibre Channel solutions have recently crossed the chasm, and that IP-based storage networks have yet to do so, it is safe to assume that the storage world we live in today will be significantly different in a brief period of time.
Historically, storage management has been a host-based process. Truly scalable, truly sustainable storage management processes, those required for the creation of a storage vision, however, can be achieved only with a centralized management strategy. One of the companies currently working to change the storage management paradigm is CreekPath.
Storage Operations Management
Scott Hansbury, Chief Marketing Officer for CreekPath, agrees that a sustainable storage management process is achieved through a centralized management strategy. CreekPath, an independent software vendor founded in 1999, has been delivering storage and SAN management software solutions since January, 2001. CreekPath currently offers an array of products designed to facilitate optimal storage operations management.
Hansbury believes that a holistic view of the IT organization and its infrastructure will prevail as the market for storage software management products matures. Products that can map critical processes and application functions to the infrastructure provide required services at every level of the storage value chain, anticipate customer demand, and will be successful in the long term.
Hansbury likens the current market for storage management software products to the market for office productivity software in the mid-to-late 1980s. Microsoft's capability to consolidate a disparate set of point products into a modular suite of solutions essentially disrupted an entire market.
Hansbury is also quick to point out that the original market predictions for office productivity software grossly underestimated the potential.
SRM is only one functional piece of storage infrastructure management, the first to see significant investment from storage software vendors. The remaining issues—policy, workflow, and automation—are currently being addressed. In addition, products that can map those functions from the application to the spindle, and are easy to install and use, will no doubt be the most successful.
Tools that provide the passive knowledge of storage resource data will eventually give way to software products that actively manage those resources. "Seeing the storage is not enough," Hansbury says. "Now it boils down to proactive management of performance and data (or information) archiving."
Point solutions on the market today address the passive and proactive pieces separately; in the near future, these products will be integrated to provide an enterprise view.
Mike Koclanes, CreekPath's Chief Technology Officer, views the evolution of storage management software as ever-expanding levels of abstraction layered on top of a guaranteed level of service, much like networking QoS (quality of service).
Koclanes believes that as IP-based storage solutions widen the choices for product deployment, the goal should be to limit the unpredictability of that service (with management techniques such as QoS). IO performance should be guaranteed to meet service levels as required, regardless of the dynamic nature of the choices available to the customer or client, and regardless of how the service levels are scoped (whether they are time-sliced, application-centric, or market-centric).
Koclanes's view eventually leads us to the need for an operations management system capable of matching available resources to required services. From a historical standpoint, the adoption rate of new, disruptive technologies is often underestimated. Koclanes believes that the adoption of centralized network storage management and the integration of Fibre Channel and IP storage network is no different. He believes that a single, integrated network that provides data services to both internal operations and external customers, and one that is capable of managing the security and billing for those services, is not too far off.
Koclanes concludes, "Software has to present itself as an application-centric networked service" to facilitate the same ease of use and interoperability in storage networking that has been achieved in computer networking.
The fact that interoperability between storage networking hardware components themselves and the software intended to manage that hardware has been insufficiently championed until now is indicative of the fact that component manufacturers have had only one goal in mind: decreasing time to market. Interoperability has been an afterthought at best with most hardware vendors, leading to dissatisfaction in the customer base. For storage networks to approach the same level of interoperability users are accustomed to seeing with LAN networking for hosts, significant energy aimed at standardization of interfaces is necessary.
Note - Fortunately for storage networking consumers, a concerted effort is underway to do just that: standardize interfaces. The Storage Networking Industry Association (SNIA) was formed in December, 1997 as a non-profit organization dedicated to the advancement of storage networking standards through vendor and end-user collaboration.
In April, 2003, SNIA formally introduced Bluefin, also known as the Storage Management Initiative Specification (SMI-S), a specification jointly developed by industry leaders in an effort to ease the pain involved with managing heterogeneous storage environments. Storage networking manufacturers, such as Qlogic, StorageTek, IBM, EMC, HP, and others, worked together for almost a year prior to the announcement for designing and finalizing the specification, which is intended to alleviate many of the interoperability and management problems brought on by a legacy of proprietary storage networking interfaces. Having the Bluefin specification in place gives vendors an open standard to code against, which ensures that end users have products that play nicely together—a luxury that has up until now been completely out of reach for storage networking consumers.
SMI-S is based on Common Information Model (CIM), another networking industry joint venture designed to take some of the headaches out of managing devices on a network. CIM is an object-oriented schema and specification that simplifies and standardizes the interfaces necessary for gathering data from objects on a network and presenting that data in a useful format.
As the first organized effort geared toward interoperability and management, SMI-S offers potentially huge gains, especially for SRM and storage management software vendors and users of those products. One of the primary stumbling blocks for getting robust storage management software to market has been the lack of standard interfaces and schemas. Historically, each HBA, device driver, FC switch, and external storage unit has been designed in a vacuum with little or no effort made toward interoperability. The majority of storage networking and storage device manufacturers has accepted SMI-S as the standard interface specification. Although it will take some time to see products that use SMI-S on the market, after the specification is ratified, consumers will start to feel some relief. Consumers will then find it easier to make the transition from an application-centric to a storage-centric entity.
An application-centric user base is more liable to see storage as cheap and disposable, and although purchase prices have fallen, the management costs for storage have yet to see significant decreases, and the TCO for storage requires constant, active management to keep in line. The knowledge gap between infrastructure and application teams with respect to the costs of storage and storage management has never been greater. As the number of terabytes managed increases, only the storage management team is aware of storage as a depreciating capital asset, and not as a cheap and disposable tool.