CIOs deploying Linux as part of their overall enterprise architecture take advantage of the superior virtualization capabilities of z/VM on IBM mainframes to enable greater agility at significantly lower Total Cost of Ownership (TCO).
Now Mainstream: Linux on System z
Once regarded as experimental or “bleeding edge” technology, running Linux on the mainframe—or Linux on System z as it’s officially known—is now a standard operating procedure in mainframe computing. According to recent Gartner estimates, approximately one-third of all System z customers now run Linux on their mainframes. Integrated Facility for Linux (IFL) shipments increased to 2,000 in 2010 and there are now more than 3 million Linux MIPS installed. Linux is continuing to grow its share of mainframe computing capacity—from 16 percent of the installed MIPS in 2009 to 19 percent in 2010 (Source: May 2011 Gartner research paper: “Linux on the Mainframe: How's It Doing, What's Next and Is It for You?”). Many large-scale financial institutions, insurance companies, and government agencies have moved well beyond proof of concept phases and now rely on Linux on System z for core processing of strategic business applications.
Start Small, Scale Up
The IBM zEnterprise 196 (z196) provides unprecedented raw computing power. Its 5.2 GHz CPU clock speed is the fastest in all of computing history to date, and a fully provisioned z196 server can deliver more than 50 Billion Instructions per Second (BIPS) (Source: http://www-03.ibm.com/systems/z/hardware/zenterprise/z196.html). You can harness that power to run new Linux applications standalone or alongside traditional z/OS workloads. The z196 can support thousands of virtualized Linux servers running under z/VM.
If you have an IBM mainframe, even an older z9 or z10 model, you probably have latent capacity you can tap to run Linux workloads simply by enabling existing hardware resources. You can configure and activate System z processors as IFL specialty engines at a fraction of the price of general CPUs. One IFL core can often support as many as 50 virtualized Linux servers (Source: http://www-03.ibm.com/systems/z/os/linux/about/index.html). So you can start small, with a few virtualized servers and scale up capacity as needed, all within the same physical frame, without any increase in server footprint or any significant increase in environmental costs.
Simplify Your Network Infrastructure With HiperSockets
Superior server virtualization is only part of the Linux on System z advantage. Using a mainframe-unique technology called HiperSockets, network connections between virtualized Linux servers or traditional z/OS logical partitions can also be virtualized via the System z hypervisor. This can eliminate cables, routers, switches, firewalls, and other network hardware, dramatically simplifying your network infrastructure and providing greater security and increased bandwidth for TCP/IP traffic between connected servers. Because data flowing over these virtualized connections never leaves the server, there’s no reason to encrypt data between HiperSocket endpoints.
Leverage Your Existing Disaster Recovery Strategy
While many customers have a mature, time-tested disaster recovery strategy for their mainframe systems, recovering applications and data on other platforms presents a significant challenge. Those same customers often find that applications deployed on Linux on System z get their disaster recovery “for free” using the same processes and procedures already in place for their traditional z/OS workloads. Linux on System z applications can typically be recovered within the same Recovery Time Objective (RTO) as their z/OS mainframe counterparts—often significantly faster than applications and servers running on other platforms.
Compare Costs Fairly
When evaluating deployment platforms for Linux, it’s easy to understand the initial start-up costs. But there can be a large difference between Total Cost of Acquisition (TCA) and TCO, which is the true cost to the business over the platform’s estimated lifecycle (e.g., three to five years).
While initial acquisition costs for distributed servers may be lower than System z, environmental costs (such as floor space, power, and cooling) often aren’t fully accounted for, especially if distributed servers are accumulated over time. In addition, “server sprawl” typically drives the need for additional support staff members, resulting in increased administrative costs when compared to highly virtualized environments such as Linux on System z.
Software licensing costs can be significantly lower on System z since Linux software is typically priced on a per-core basis. So, depending on individual product licensing terms, replacing 50 x86 cores with virtualized servers on one IFL could reduce the cost of software running on the servers by a whopping 98 percent.
Leverage Management Tools
You don’t have to take the plunge to mainframe Linux alone. Several vendors offer a variety of tools to help your technical staff manage both z/VM and Linux on System z. These tools can help in areas such as server provisioning, system monitoring, problem triage and resolution, chargeback, capacity planning, system automation, data protection, backup, and restore. In addition to improving the productivity of your technical staff, these tools can help drive higher application availability and reduce recovery time when problems arise.
In today’s difficult economic times, the pressure is greater than ever to gain a competitive advantage by lowering costs while continuing to deliver high qualities of service. Smart CIOs are looking to their technology leaders to maximize the use of Linux in their strategic architecture. They’re finding that running Linux on their System z mainframes delivers lower costs while increasing reliability, security, and availability, yielding a significant competitive advantage to their critical lines of business that rely on IT.