Emerging Storage Technologies: Promises, Promises
Prior to the economic downturn during 2001/02, there were no fewer than 299 startup storage software management vendors. Although many of them have since been purchased by larger vendors or ceased doing business, this market category remains dynamic, with numerous promising technologies from both established and emerging vendors.
META Trend: Through 2006, the lack of robust storage management automation, standards, and process maturity will limit net annual enterprise storage capacity growth to 45% (58%-67% gross procurement), with price/capacity improving 35% per year. To effectively leverage and manage enterprise media assets, users will require a data/media center of excellence. Through 2007, software value-added functionality, manageability, integration, and interoperability will be the primary enterprise storage differentiators, rendering storage hardware a tiered commodity.
Open systems storage management as a discipline is currently in its adolescence. Few collegiate computer science programs offer storage management-specific courses. As a result, most IT organizations (ITOs) must either develop storage management skills internally or hire them from other organizations. Given that most other IT disciplines have higher levels of maturity and therefore available labor pools (e.g., servers/systems management, network management, database management), experienced storage managers will be in disproportionately high demand through 2007/08.
The storage management software and appliance markets (i.e., storage-specific software hosted on servers) will remain fragmented through 2005/06. However, given the buying binge of major vendors (e.g., Veritas, EMC, IBM, HP) acquiring startup companies during 2001-03, we expect these vendors to begin offering integrated suites by 2004/05. Indeed, all major vendors are currently making progress in this regard. As a result of improved tools, storage personnel productivity will increase 10%-20% by 2006/07.
Intelligent Switches
Although dominated by established vendors, the intelligent switch category is among the most pivotal and important emerging technologies. Intelligent switches will host many storage applications (e.g., virtualization, data movement, provisioning, security), thereby complementing and, in some cases, eliminating the need for many special-purpose storage appliances (e.g., fabric security, in-band virtualization) by 2005/06. Such special-purpose appliances are generally offered by each individual vendor to assure a platform on which to host its applications (e.g., IBM SAN Volume Controller, HP CASA/VersaStor, DataCore SANSymphony, FalconStor IPStor, StoreAge SVM). At the same time, new special-purpose appliances are emerging for applications such as security/encryption and storage resource management.
This proliferation of devices has the potential to significantly clutter the fabric, create conflict between devices, and consume numerous switch ports, thereby increasing the cost of the solution. Moreover, an array of disparate devices complicates security, especially because interdevice standards are not likely to be widely adopted before 2005/06.
Vendors/products contending for this market include Brocade Fabric Application Platform (formerly Rhapsody), Cisco’s MDS 9000 family (formerly Andiamo), McDATA (via its August 2003 acquisition of Sanera DS 10000 and its strategic investment in Aarohi technology), and MaXXan. At a time when switches are becoming increasingly less differentiable, competition is increasing, and prices are under intense pressure, vendors are eager to introduce highly distinguishable products that offer a sustainable competitive advantage.
As with the primary switch market, intelligent switches will be sold primarily by original equipment manufacturers (OEMs; e.g., IBM, HP, EMC). There are relatively small numbers of major OEMs, and attracting them is vital to the success of any intelligent switch. However, given the disparate architectures of these devices, the engineering effort by OEMs to support each device is significant. At the same time, OEMs will be reluctant to commit to a single vendor device or alienate themselves from any significant market segment. Thus, OEMs will likely commit to at least two but probably no more than three devices through 2005.
Business Rule Definition and Process Development
META Group defines storage virtualization as “the ability to manage a heterogeneous pool of storage as single entity by policy”. With the maturation of storage virtualization by 2005/06, the critical component to fully realizing the potential of virtualization will be establishing comprehensive business rules, process management, and related storage policies. These policies and rules will be key to implementing life-cycle data management (ILM). Although numerous vendors are currently touting ILM, many such offerings are little more than new product packaging.
We estimate that fewer than 25% of ITOs) have established and documented storage processes and policies. Without such forethought and documentation, it is impossible to implement automated practices. ITOs that do not feel as though they have sufficient expertise to perform process automation and documentation should consider engaging a third party to assist in the effort. Although major vendors are available (e.g., IBM, HP, Accenture), smaller specialty organizations (e.g., Glasshouse Technologies) are also in the running. In many cases, the benefits of such a project may include significant improvements in personnel efficiency without any additional automation.
Storage Encryption and Security
Although fabric security is available from switch vendors (e.g., Brocade Secure Fabric OS, McDATA SANtegrity Security Suite), it is limited to issues such as device authentication, command encryption, and hard zoning. However, these products do not encrypt data either in the data path or on the hard drive. For this application, emerging vendors/products include Kasten-Chase Assurancy and Decru DataFort. Given recent regulatory requirements regarding data confidentiality (e.g., HIPAA in the US), data encryption will become a standard practice in high-risk organizations (e.g., healthcare, financial services) by 2004/05, and we expect switch vendors to acquire these emerging vendors or others like them.
As previously mentioned, applications such as Assurancy and DataFort require their own fabric appliances to encrypt and decrypt data. Having such an appliance can complicate fabric architecture and data path planning and increase the number of components that can possibly fail in the fabric. Therefore, intelligent switches become an ideal hosting platform for this functionality and are likely do so by 2005/06.
Business Impact: Because of current product innovations, many highly promising new technologies are being offered by both established and emerging vendors. ITOs that examine both may find valuable diamonds in the rough.
Bottom Line: The storage management market is in a period of rapid product innovation. ITOs should first establish/improve their business processes and then seek products to automate them.
META Group originally published this article on 16 October 2003.