Mainframe lessons: slowly moving to non-mainframe systems.

Storage demand grows faster than the effective deployment of management tools, and the supply of trained people to manage storage has fallen behind the demand. Have you ever heard comments to this effect before? The IT community worldwide is anxiously awaiting the point in time when it is perceived to be both easier and more cost-effective to implement a comprehensive storage management than just adding more hardware. That day may not be that far off after all.

Effectively managing storage involves far more than ensuring enough capacity is available to meet increasing demand. In reality, that's the easy part. Choosing between a broad range of management tools from hundreds of vendors, selecting from a long list of backup/recovery, data security, and virtualization products has become too time consuming and complex for most businesses today given their limited resources. Disk and tape technologies from multiple vendors, network management, and identifying SAN, NAS, and DAS tradeoffs add to the time-consuming challenges. In addition, storage management capabilities vary widely by operating system. As a result, delivering the levels of service and availability required by increasingly critical business applications has been pushing the reality of effective storage management farther away for most non-mainframe businesses.

The First Storage Management Architectures Arrive

Storage management functionality didn't initially extend far beyond backup and recovery for disk systems and the highly effective TMS (Tape Management System) for mainframe tape systems. By the early 1970's TMS was the popular de-facto standard tape management system and it is now part of CA's (Computer Associates) storage management product family, called CA-1 Tape Management. CA-1 provided management with the capability to control and protect tape data sets and volumes. It was the first tape management system to enable tape library inventory tracking with a complete audit trail; tracking of off-site vaults; standard and customizable reporting; support for pooling to ensure scratch tape availability; and utilities for controlling tape and catalog maintenance activities.

Storage management for disk storage was popularized by IBM's HSM functionality beginning in 1976 and was soon followed by DSM/OS, an HSM product from Sterling Software that quickly gained appeal for its performance, ease-of-use and policy-based capabilities. Via the acquisition of Sterling Software, DMS/OS, now called CA-Disk, is part of CA's mainframe storage management offerings.

In April, 1988 IBM announced DFSMS (Data Facility Storage Management Subsystem) for their MVS (mainframe) computers. Now commonly called SMS, this architecture consisted of a set of related software products that marked the most comprehensive set of storage management capabilities introduced to that point of time. The SMS framework provided a policy-based storage management solution for large mainframe computer systems and is an integral part of OS/390 and currently z/OS systems. Its primary goal was to provide policies and automate the most significant tasks of data storage administration.

In time, SMS became an effective policy engine for managing storage resources and required users to get to know their data and to better understand its value enabling businesses to have access to the right data at the right place at the right time. There are several components within SMS, but the well-established HSM (Hierarchical Storage Management) functionality was the catalyst that enabled businesses to ultimately address the storage capacity dilemma of matching policy based data attributes with the most cost-effective storage technologies. Today, CA-Disk and DF/HSM are the only two products providing true policy-based HSM functionality for mainframe systems.

Though SMS was definitely not an end-to-end storage management architecture, SMS represented the most comprehensive achievement in storage management software to date and it became a mainframe standard for storage management after nearly 10 years of evolution. In parallel, relatively few innovative mainframe storage software companies continued to enhance SMS or its individual components. Many of the remaining requirements are being met by these companies enhancing its value while enabling effective storage management on the mainframe to become a reality.

Storage Management for Non-Mainframe Systems

What about SMS-like capabilities for today's typical IT environment that deploys multiple heterogeneous operating systems such as Unix, Linux, Netware, Windows and mainframes? One of the biggest remaining unaddressed issues for SMS and a comprehensive storage management solution is that it's only available on mainframes and is not extensible to any other computing platforms. A few savvy storage management vendors are addressing the need and significant business opportunity for an enterprise-wide policy-based storage management solution. This is increasingly important since the Unix and Windows (non-mainframe) systems now generate more than 85% of the world's stored digital data (whitch clearly justifies a legitimate, cross-platform storage management solution. Many non-mainframe businesses demand the mainframe-class storage management functionality that they have counted on for over 15 years. Today's biggest storage management problems are centered in these non-mainframe systems.

Throughout the 1990s, the easiest way to manage storage for non-mainframe systems had traditionally been accomplished by simply adding more storage capacity. As the management gap between storage capacities, the ever-increasing value of digital data, and the number of trained storage administrators continues to diverge, this strategy is no longer effective.

The Management Gap Widens

We recognize that installed storage capacity for non-mainframe systems continues to grow faster than the deployment of management tools, and the supply of trained people to manage storage has fallen behind the demand. Aligning data and storage to the business has become one of the IT industry's fastest growing and most critical requirements. By 2008, it is expected that the "average" non-mainframe storage administrator will be able to manage nearly 20 terabytes of storage while the amount of data to be managed will approximate 80 terabytes. Because adding more administrators is seldom an option, several steps can be taken to help close the management gap. This gap is referred to as the storage management gap or "The Infinite Disruption".

Today, the typical mainframe storage administrator manages on the average more than 40 terabytes of disk storage while the typical storage manager for Unix, Linux and Windows manages from 1 to 5 terabytes. Remember these are averages and can vary considerably. It has now become a primary goal of storage administrators using these systems to achieve the same high-level storage management capabilities available on the mainframe. Until recently, if you wanted to manage all storage effectively with robust tools and intelligent policy-based software, then moving to a mainframe was the only solution available.

[FIGURE 1 OMITTED]

Non-Mainframe Systems Issues Mount

Why does a storage administrator on a mainframe manage, on the average, well in excess of 40 terabytes of online storage while non-mainframe storage administrators manage just a few terabytes in comparison? The single requirement for a centralized, end-to-end storage management solution using a "single pane of glass" for mainframe and non-mainframe systems that really works across multiple operating systems could not be greater!

Beyond mainframes, the Unix operating systems host more critical, data intensive applications than any other operating system. The storage services provided by Unix that are present today are basically the same ones that existed over 20 years ago when Sun first released NFS (Network File System) and have seen little improvement since. There have been no significant enhancements to non-mainframe operating systems in terms of storage services since those early days when a "large" Unix server had about 1 gigabyte of disk storage and PC's were just standalone boxes. No one knew then that these distributed computing systems would one day be asked to do the work of a mainframe and to access as much, if not more data than their mainframe counterparts.

[FIGURE 2 OMITTED]

Meeting the Challenge ...

The non-mainframe systems were originally designed to be computational systems, not managers of large storage pools as they are today. If much foresight had existed, the limited storage services for these systems would have been viewed as unacceptable and would have been addressed long before now. Multiple versions of Unix emerged and each adopted their own often incompatible file systems, volume managers, backup/recovery/security products and data movers making interoperability between unlike platforms difficult to achieve.

By the end of 1990's hundreds of independent storage management software products were developed attempting to address the increasing number of mainframe and non-mainframe storage management problems. There is little basis to believe that non-mainframe operating systems can be scaled to the level necessary to replace the mainframe class servers without having the same type of tools available that made their legendary predecessor so successful. Industry consolidation may offer more immediate help. For example, the Sun acquisition of StorageTek in Aug. 2005 creates a complete hierarchy of hardware solutions that can be bundled with an expansive suite of storage management software and data protection capabilities.

Missing Pieces for Non-Mainframe Systems

Some pieces were still missing however as the 21st century began. Deployment of end-to-end, cross-platform management solutions providing functionality such as SRM (Storage Resource Manager) and SAN (Storage Area Network) virtualization capabilities remained scarce. Nonetheless, the looming challenge for non-mainframe systems remains much larger given the tremendous amount of data and higher growth rates that exists on these systems.

The critical list of storage management requirements for IT organizations includes:

* A centralized, single console integrating SRM (Storage Resource Management) and all related storage management and reporting capabilities for mainframe and non-mainframe systems (a common look and feel)

* More effective capacity utilization of tape and disk resources

* A single, effective, proactive HSM for heterogeneous non-mainframe systems

* An easy to implement method to shrink backup window requirements using new technologies such as Commonality Factoring

* A way to improve storage management productivity while minimizing staff impact

* A knowledge base to log, trend and solve problems from current and historical data

* Improved business continuity by providing alerts with policy-based actions (pro-active management) to address problems before they become an issue

* Vastly improved levels of data security/protection to prevent intrusion, worms, viruses, scams, phish, spyware (data theft), and the growing threats of "digital diseases," etc.

The company that best understands and addresses these critical issues will be ideally positioned to become the industry leader for storage management solutions.

Overcoming the Barriers--SMS Capabilities for Non-mainframe Systems

There are numerous software companies providing solutions to various storage management problems for non-mainframe systems. The sheer number of product choices makes buying decisions complicated and it is impossible for all the various products to interoperate adding to the interoperability challenge. This suggests that a consolidation of these vendors will continue, and the goal of true Open Systems interoperability will continue to be a dream versus a reality for some time. Multiple HSM-like products exist for open systems but none readily support the three fundamental Open Systems computing platforms; Unix, Windows, Linux and the range of databases and file systems that that are normally deployed. Therefore a business often needs to install multiple, semi-compatible, or non-compatible software products to attempt to manage their storage environment. This complicated approach is exactly what businesses no longer want or can afford to do. Coupled with steadily increasing storage growth, this represents an unsustainable strategy. Fortunately some up and coming software companies are beginning to address this longstanding opportunity.

[FIGURE 3 OMITTED]

Conclusion

The era of combining weak operating system storage management services with very large storage subsystems is rapidly coming to a close as leading-edge ISVs are intensifying efforts to resolve this mounting problem area and are starting to bring mainframe-class capabilities to the rest of the IT community. The storage industry is now embarking on a new course leaving behind many of the tenants that defined the way the non-mainframe storage management business was conducted for the last 25-30 years. Expect cutting-edge technologies that have the potential to offset the losses in the old markets to move into, and create new markets, and possibly new industries, as the value of data skyrockets. Storage management will definitely be one of them.

Fred Moore is President of Horison, Inc. (Boulder, CO).

www.horison.com