zStorage Industry Update and Trends zSTORAGE

2 Pages

Some storage trends can be fairly predictable and consistent. For example, more data is being stored with more copies being made and retained for longer periods. Likewise, it’s safe to say storage capacity, performance, and features will continue to advance. Recent events that will continue in 2006 include:

  • Continued vendor consolidation
  • Larger capacity, higher performance, more reliable, and physically smaller devices
  • Focus on data protection, retention, and regulatory compliance
  • Data and information security (e.g., encryption, asset tracking, and media disposal)
  • Interest in data classification and policy-based automated storage management
  • Pursuit of interoperability across hardware and software
  • Automated conversion from smaller mainframe volumes to larger capacity volumes
  • Growth of distributed and remote replicated data as well as network bandwidth
  • Continued interest in storage virtualization, Information Lifecycle Management (ILM), grids
  • Continued interest in fiber-based access using Network-Attached Storage (NAS)
  • Wide Area Network (WAN) Services and Wide Area File Services (WAFS) awareness.

Highlights from 2005 and early 2006 include:

  • Sun acquired StorageTek (STK), Seagate is buying Maxtor, and Imation is buying Memorex
  • Hewlett-Packard (HP) acquired storage and management software vendors AppIQ and Peregrine Systems
  • EMC’s acquisitions, including storage virtualization vendor Rainfinity and pieces of defunct switch start-up Marranti
  • IBM refocused on NAS with a partnership with Network Appliance, which bought storage security startup Decru
  • Symantec bought open systems storage software vendor VERITAS
  • Cisco Systems made several acquisitions, including InfiniBand vendor Topspin Communications
  • Brocade invested in WAFS and application accelerator vendor Tacit Networks
  • Legacy storage networking and mainframe connectivity vendor McData acquired CNT, which had just digested its 2003 acquisition of INRANGE.

Compared to the open systems environment, the number of new vendors in the mainframe storage market is rather small. However, from a storage perspective, new development is still taking place for mainframes. Specifically, last year:

  • EMC introduced the latest Symmetrix DMX-3 that now supports 2,400 disk drives
  • Softek added an LDMF data migration facility
  • Sun Microsystems (formerly STK) offered enhanced disk and tape solutions
  • Hitachi Data Systems (HDS) provided TagmaStore USP and new NSC models
  • IBM made processor, disk, tape, and software enhancements as well as released FICON Express2 for z990 and z890 systems and a statement of direction that the SAN Volume Controller (SVC) will gain zSeries support in the future but only for FCP access by Linux images running on zSeries.

For Storage Resource Management (SRM), Microsoft Excel remains a popular tool in non-mainframe environments. Excel doesn’t do the data collection, modeling, and other advanced functions found in traditional performance and capacity planning or SRM-type tools, but it’s good enough for many environments. Some vendors have figured out that by linking SRM to traditional server and storage performance and capacity planning, they spend less time on missionary work (i.e., educating customers about the importance of resource monitoring) and more time discussing solutions.

Data classification is another storage topic to keep an eye on. Solutions for early adopters of data classification technology are emerging. Deep data classification technology remains in its infancy, especially for large, mission-critical production environments. Learn and understand the differences between content and context-based classification and how it applies to you.

Data differencing, also known as data de-duplication, is a technique to eliminate duplicate data; it shouldn’t be confused with compression. Examples of technologies using data differencing include WAFS and WANS bandwidth optimization solutions and technologies such as backup, replication, and mirroring disk-based backup libraries. By identifying duplicate files, only one copy of a file must be saved; this reduces the capacity needed. Another example is to reduce the amount of duplicate data blocks that must be sent over a network link; this improves latency and optimizes bandwidth.

Confusion lingers as to what is and isn’t grid as well as what is a service, product, or architecture approach. Will 2006 be the year of the grid? That depends on your interpretation of what a grid is. It’s similar to ILM; both terms are used liberally to refer to different things. Grids are being used to refer to on-demand services, including remote storage capacity for sale or rent, managed services, compute clusters and server farms, storage systems composed of clustered general-purpose processors and many others.

Ask yourself a few questions:

  • What is the real benefit of a grid solution? What is your storage access profile?
  • Do you really need a grid or want one only because it’s the latest technology?
  • Which of your applications need a grid? Is the grid for compute or storage purposes?
  • How would delivery of storage services to your environment improve with a grid offering compared to traditional approaches?

Data protection, security, and business continuance should remain popular issues for the foreseeable future, given increased media coverage and awareness of information and data loss as well as regulatory and data privacy pressures.

To address security requirements, the focus has been on increased use of encryption and reducing the physical movement and handling of data. Key management is an important aspect of encryption. If you encrypt your archives, will you be able to access and unlock your encrypted data in several years? How will you manage the keys and tools to encrypt and decrypt data? Where to encrypt is another question; should it occur while the data is at rest on storage media, while in transit, or via a combination of approaches?

2 Pages