Since its introduction, the IBM zEnterprise System—the combination of mainframe central processing with zEnterprise Unified Resource Manager (zManager) and zEnterprise BladeCenter Extension (zBX)—has stirred significant interest. At the time of launch, some questioned whether a real market existed for the zBX-mainframe combination. A little over a year later, the answer is definitively positive. CIOs at more than 100 IBM clients (as of Jan. 26, 2012), in multiple industries around the world, use the zBX to solve a range of business problems. Learn how they’re benefiting.
The three main elements of the IBM zEnterprise System appear in Figure 1. At the heart of the system are zEnterprise mainframe servers, including the z196 Enterprise Class (EC) and z114 Business Class (BC) servers. Each mainframe can be connected to a zBX, which integrates IBM POWER7 and IBM System x blades, and one or more specialty optimizers (e.g., IBM’s DataPower XI50z) can be added. Finally, up to eight zEnterprise/zBX nodes can be managed as one logical system.
zManager software manages all supported system platforms and operating environments. From a single point of control, it provides installation, monitoring, management, optimization, and diagnostic services for zEnterprise resources and workloads.
The zBX server integrates IBM POWER7 and IBM System x blades, letting them share resources and operate diverse workloads with the mainframe while being managed from a single system. Two private networks, a 10GB Intraensemble Data Network (IEDN), and a 1GB private Intranode Management Network (INMN), connect the zBX and the zEnterprise Central Processing Complex (CPC). These private, internal networks reduce the cost and complexity of connecting the platforms. The results can reduce labor, management, and network infrastructure costs by as much as 90 percent.
Here’s what CIOs are finding so attractive about the zBX:
10. One user interface to rule them all. zManager consolidates management and administration tasks into a single, consistent interface while automating task processes across all three zEnterprise platforms (System z mainframes, System p distributed systems, and x86). Administrators manage and configure servers, network and other resources (e.g., specialty optimizers). Cross-platform consistency eliminates the need to have different administrative support and skills for each platform, which simplifies management and reduces training costs.
9. Automation cuts costs. Automation replaces many manual processes in five critical management areas: asset, deployment, capacity/performance, security and change for Systems z, x and p. Reducing installation and configuration errors lowers operational and labor costs. IBM’s Competitive Project Office (CPO), which works with IBM customers to analyze solutions, focusing on factors that are most important to that customer, has released measurements and projections showing aggregate labor costs can be reduced by more than 30 percent.
8. Wizards create networks faster and cheaper. zManager wizards automatically define, configure, assign, and operate VLANS across platforms. Servers on a VLAN can only “see” servers connected to the same VLAN. Multiple VLANs mean virtualized Research & Development (R&D) servers can operate securely and privately as they’re isolated from production servers running on the same platform. Network capital and operational costs are lower as external router acquisition and management costs are eliminated.
7. Over-provision no more! The costs and frustration of hardware over-provisioning are reduced by automated performance management for z/VM and Power virtual servers (extending to x series in 2012). CPO measurements show such management can reduce hardware over-provisioning by as much as 10 percent, though the actual amount depends on the workload. Costs are lowered by alleviating the need to buy hardware and operating with fewer support staff.
6. Fewer cores means lower software licensing costs. Distributed software pricing is typically tied to the number of cores; less hardware means fewer cores. Performance management reduces hardware over-provisioning, resulting in a lower processor core count and lower software costs.
5. Improved security eliminates firewall cost. Consolidating distributed servers into zBX moves server-to-server traffic into VLANs on the private IEDN and off exposed external networks. This reduces, or may eliminate, the need for firewalls between zBX servers, as well as between zBX servers and those hosted in attached zEnterprise CPC(s), avoiding the cost of installing, maintaining, and administering those firewalls.
4. Consolidation vanquishes complexity. Consolidating distributed servers into zBX and Linux on System z virtual servers eliminates more than 90 percent of network infrastructure components (cables, switches, and routers). The cabling between the zBX and attached CPC(s) can be reduced to as few as 20 network parts. zBX consolidates Systems p, x, and special-purpose devices into a single, integrated chassis with shared access to power and networks. This simplifies access to and management of power, cooling, network, and floor space, while significantly reducing operational costs. zManager can set, manage, and control hardware energy caps and control operation to minimize energy consumption. The result is dramatically reduced maintenance, support, and implementation costs as well as lower power consumption, management, and administration costs.
3. Wizards vaporize virtualization costs. All servers running under the zManager are virtualized. Automated wizards define and create servers in minutes. Duplicating servers for test, development, or to take up operational load, occurs quickly, reliably, and without error, using either existing server definitions or customized definitions. This reduces the costs of hardware and administration commissioning/decommissioning and increases infrastructure utilization.
2. Reliability, Availability, and Serviceability (RAS) reduce downtime. zBX is a blade center designed for mainframe connectivity and RAS performance. The zBX includes such original blade center design features as increased power and data resiliency, minimized floor space, hot swappable power supplies and higher computational density. Blade centers deliver equivalent computing power in half the floor space needed by rack-mounted servers. This puts twice the computing power in the same physical space; when compared to standalone/tower-based servers, the reduction is even more impressive. Dual data connections and dual power supplies ensure neither will be a single point of failure. Continuous self-monitoring and hot swappable features ensure that if a zBX power supply reports a potential problem, it can be replaced without bringing down the zBX or any servers running in the zBX.
1. Single-source support means only one throat to choke! Once IBM finishes installing the zBX, blades and special-purpose hardware, support and maintenance services are provided as part of the mainframe. The zBX maintenance procedures follow the same processes used to service other System z family members. Consolidated, consistent processes, combined with a single point of contact and control for support and maintenance, reduces costs, improves services, and lowers administrative complexity. Given the variety of platforms (System z, POWER, and x86) and operating systems (z/OS, z/VM, z/VSE, z/TPF, Linux, UNIX, AIX, and Windows), this represents a great value to the user.
As noted in the sidebar, the zEnterprise System and zBX are delivering value and much-needed savings to CIOs. Future articles will deal with specific operational areas where zBX customers are accumulating these savings.