Meta trend
By 2006, OS/390-z/OS's onerous software costs (2x-5x those of high-end Unix and about 10x those of Windows 2000) and 15 to 20 percent annual hardware price/performance improvement will slow the annual net capacity growth of S/390-zSeries to less than 15 percent (versus Windows 2000 at more than 60 percent and Unix at 50 percent). By 2011, the capacity composition of enterprise data centers will approach 48 percent Windows 2000, 42 percent Unix/Linux, and 10 percent OS/390-z/OS (legacy).

[an error occurred while processing this directive]

An analysis of hardware partitioning (a.k.a. virtualization) techniques, workload management (WLM), and availability remedies for high-end Unix suppliers (Sun, HP, and IBM) indicates that, in the past 12 months, Unix vendors have turned the corner in their pursuit of mainframe-like systems and mission-critical delivery. Although historical Unix software and hardware virtualization maturation curves have been often just "technically interesting" (for mainframe operations), Unix virtualization options are now capable of delivering more high-end delivery options. Major Unix vendors now have hardware as well as software critical mass, and our extrapolation of each vendor's two to three major operating software and hardware upgrades (about every 18-20 months) indicates that all three should reach near parity with z/OS (as we currently know it) in the 2005-07 time frame. Although high-end systems will tend more toward 2005, middle-range systems will have to wait until 2007 for these upgrades.

In fact, by 2005, all will provide dynamic, logical partition virtualization (i.e., not tied to hardware board boundaries a la Sun) via new operating firmware software virtualization layers. In addition, all major Unix vendors will provide fine-grain WLM, leveraging emerging fair-share processor allocation schemes that will incorporate swap data set management shortly (2002/03). Further, important goal-driven WLM extensions with application group-like aggregation (much like z/OS) will emerge within 24 months (with reasonable maturity, even by z/OS standards, by 2007). In fact, in the next five years (from a z/OS perspective), these Unix vendors will move the Unix platform forward almost 15 "z/OS years" (1990 to 2005) because little discovery work is required for expanded availability, WLM, and partitioning frameworks. Because so much has already been invented and concepts proven, for high-end Unix suppliers, it is a case of planning the work and working the plan as they replicate proven technology concepts.

However, z/OS is not standing still. For example, its Intelligent Resource Director (next-generation, expanded WLM), which monitors and manages clustered logical partition performance with dynamic I/O reconfiguration management, will continue to evolve, providing support for non-z/OS images and, in the 2004/05 time frame, these same services across Sysplex clusters. By 2005, z/OS's already strong I/O subsystem will have received the major overhaul needed to minimize many of the now non-parallel functions that will become a performance bottleneck as processor speeds continue to increase. Although z/OS is a moving target, a 2002/03 version of z/OS-like functions mapped to 2007 Unix hardware will prove very enticing to most large shops.

1 2	| Next page

Newsletters
	Tech Update Today eBusiness Update	Tech Update Weekly		All newsletters FAQ Manage my newsletters

[an error occurred while processing this directive]

[an error occurred while processing this directive]