My first foray into the `mainframe’ world was when I began working on Linux solutions for z back in 2001. It was fascinating to see how many people had a bias against the platform, whether they knew anything about it or not. I tried to focus on educating people, and showing them the clichés around old mainframes no longer applied. Reactions were mixed – some had open minds, others did not, and some were inexplicably hostile to the platform. Whatever the reaction, I usually found at its root the same deep passion for the computing world that I first discovered back in 1980. Like anything, this passion can be used to drive one down many paths. For some it opens their minds to considering new options, new directions and new strategies. For others, it drives them more deeply into supporting their technology of choice.
Today’s z Systems are open, flexible, and competitively priced based on value. In fact, today’s z Systems are the most flexible systems on the market, and every data center can benefit from flexibility.
What do I mean by flexibility? It is the ability to meet varied scalability and architectural requirements with minimal disruption – without compromising performance or security, and without adding significant cost. I/O rate is a common weakness of Intel based systems. For example, we often see Intel systems where I/O is a bottleneck running at utilization rates around 10% for non virtualized systems, and 50% for virtualized environments. This translates to more cores needed to meet SLAs, and software costs begin to skyrocket. Conversely, z Systems can meet SLAs using fewer cores, and run at well over 95% utilization comfortably with high sustained I/O rates. If you think about it, just combining these two factors provides a significant advantage for z Systems.
That said, many of the z Systems flexibility features can be found in other systems. And although some systems handle certain tasks better than z Systems, when you take all the features into consideration as a whole, z Systems begins to shine quickly.
Let’s see why. Here’s a list of advantages z Systems offers to support my case:
- z Systems can host seven different operating systems.
- z Systems provide the most secure and discreet LPAR technology available.
- HiperSockets allow secure, high speed connections between LPARs, including Linux LPARS connecting to z/OS or z/VM VSE LPARs.
- Private, public and hybrid clouds can be hosted on the same physical machine, and accommodate the differing data, front end, security, high availability, and disaster requirements each might have.
- The 141 processors available on a z13 can be characterized to fulfill multiple roles.
- Robust internal level 2 and 3 networks with VSwitch and Guest LAN options. Network traffic can be kept inside the box, or routed out and through firewalls before returning.
- PRISM and z/VM Workload Manager provide advanced resource assignment and policy based workload management.
- System Assist Processors (SAP) provide I/O instruction execution and off loading to a degree that allows significant reduction of workload run times and performance increases.
- CPU, memory and I/O scaling, LPAR creation all performed without disrupting the system.
- Disaster recovery for both the z/OS side of the z System and the Linux side can be integrated or kept separate.
- Mobile application servers running on Linux can benefit from the strengths of z Systems and provide direct, fast connectivity to data and processes hosted on z/OS and z/VM VSE – all while still being integrated with advanced z hosted security and other systems hosted on other platforms.
The unique combination of running z/OS and Linux on the same box provides us with many unique opportunities, and its unmatched security allows us to do things that otherwise would not be possible. Its system configurability also allows us to mold a system to exactly what a client needs.
What does this mean to you? You can have a system that can run mission critical databases, as well as backup services, DNS, message queuing, Web application servers, systems monitoring, and asset management. When a certain application is not needed, it is paged out and consumes no resources until it is needed again. Workload management ensures applications are given the resources they need. On the fly scalability is available exactly when you need it. You have a system that can be useful not only for `mainframe’ applications, but for many other applications. Fewer cores, same performance, lower software costs. Isn’t this something you should be taking advantage of?