|System z on Facebook
The 12 Days of z-mas
Caroline Exum 270004MPQK firstname.lastname@example.org | | Tags:  disaster recover z" reliability efficiency zlinux costs linuz "system z mike_baskey licencing dr tco system security zenterprise openstack scalability workload management open | 3,254 Visits
Linux was brought to the mainframe over 12 years, started by a college intern. When I heard about the project, while most of my friends laughed, I was intrigued. What if our existing mainframe customers could take Linux applications and run them on the mainframe and benefit from our virtualization and reliability. Add hipersockets to enhance the ability to access existing mainframe applications and data and you might have something. The rest is history.
Let's take a closer look at the top 12 reasons for running workloads on System z – we'll cover each of them in more details in coming weeks/months.
The first 3 are the typically cited ones – 1). software license costs (due to the incredibly powerful z196 and now zEC12 systems), 2). energy efficiency and 3). large scale virtualization (including TCO- see below). But there are many more.
Fourth, there's workload management. Our clients typically run in the 80% CPU range, which is 2-3x more than what analysts recommend on distributed (x86) servers.
Fifth, there's disaster recovery. Look at the true cost of providing this on distributed systems and compare, including all the configuration required. Why risk your business?
Sixth, there's security. Consider the cost of not only complying with audits and proving your data is protected and isolated but also the cost of being found in non-compliance. Advantage Z.
Seventh, there's reliability. How much time do you spend after an outage in post mortem calls with the typical finger pointing exercises from all the distributed teams. (see the Sicoob video www.youtube.com/watch?v=E8cdYINr32M). Z hardware mean-time-to-failure measured in years (double digits).
Eighth, there's scalability, especially I/O bandwidth and CPU on Demand. Private clouds need capacity “instantly” - compare the Z model with procuring/re-provisioning x86 servers.
Ninth, there's the data center footprint. Many enterprises and institutions have limited space but have need to grow as they expand services and leverage their infrastructure.
Tenth, consider the overall TCO. Look at the ability to maintain your data center with fewer people on a unit cost that scales almost infinitely. Hard to ignore.
Eleventh is open. Starting with OpenStack in z/VM (with XCAT inside). Much more to come.
Twelfth and most importantly are the numerous client examples where they're seeing real value in their every day business operations– check out some client stories from: Nationwide, Algar Telecom, Blue Cross and Blue Shield of Minnesota, German Pension Fund, and Bank of Tokyo-Mitsubishi UFJ.
So, what the heck are you waiting for!
Find out more at http://www.ibm.com/systems/zlinux
Mike Baskey is an IBM Distinguished Engineer and member of the IBM Academy of Technology. He served as chairman of the DMTF standards board from 2007 to 2011 and is now the Chief Architect for Management for System z with specific focus on cloud, IT analytics and mobile.