IBM Z Hardware and Software Maximize Resiliency
In a perfect world, systems should always be available. And we have a computing system with that capability: IBM Z.
By Bill Seubert and Stephan Wiedemer07/12/2019
What’s “resiliency?” One dictionary defines it as “the capacity to recover quickly from difficulties; toughness” or “the ability of a substance or object to spring back into shape; elasticity.” When we think of the resiliency of computer systems, we also think in these terms. If a system fails, how does it “spring back?” How quickly can it recover? How tough is that system?
But it’s not enough to simply allow the system to spring back. It’s far better to prevent it from having to spring back in the first place. In a perfect world, systems should always be available and have the capability to avoid situations that would allow them to fail. And we have a computing system with that capability: IBM Z*.
Customers Expect Resiliency
Resiliency of computer systems can become a critical issue. For example, customers waiting to do business transactions while your system is down will likely turn away to another vendor. Often, an alternative for the product or the service you’re offering can be found with a couple clicks. The odds are high for missed opportunities and lost customers.
In an economy of speed, business disruptions can become fatal to an enterprise. Your company can face immediate threats to your global brand by dissatisfied customers complaining about your service quality on social media. At the very least, your brand value and reputation can be degraded or even lost.
Because we live in a 24-7-365 digital economy, resiliency is a key business imperative for customers. If firms don’t deliver the seamless experience customers expect—regardless of what causes an outage—it creates an unforgiving situation.
DR and CA
Resiliency can be looked at in several ways. For many years, companies have focused on how to handle disasters with disaster recovery (DR). DR plans focus on continuing operations, identifying required backup data, establishing alternate sites, examining the geographical characteristics required for a backup site, calculating how much data loss and downtime a company can tolerate, and more.
The ultimate goal is to have systems with continuous availability (CA)—those that are always on and available, at least within the bounds of the requirements of the business. Systems that deliver CA must be available no matter what planned or unplanned outage might occur.
IBM Z Provides CA
IBM strives to provide CA for the IBM Z mainframe system, to the point where IBM named mainframe systems “Z” to represent “zero down time.” The IBM Z server has been known for years as a “five nines” system—one that can achieve 99.999% availability. Many hardware and software features make this possible. Some are inherent to the hardware, OSes and other system software. Others must be implemented by clients apart from already built-in functionality.
IBM Z processor and storage hardware offer a vast number of functions and features that are focused on providing high availability (HA). For example, processor cores are dedicated as spares that can be used if another core fails. IBM Z memory uses a redundant array of independent memory to protect from memory failures. And multiple I/O device channels can be configured to provide HA and high performance. Dozens of other reliability, availability and serviceability features are included in IBM Z hardware.
z/OS* HA functionality starts with OS routines that help failing components recover. The most prominent feature of z/OS for HA is Parallel Sysplex. This is the clustering function for z/OS. A Parallel Sysplex can cluster up to 32 z/OS systems, interconnected by a shared memory feature called the coupling facility. Systems communicate via ultra-high-speed connectors between the systems, and also use a shared time synchronization feature. Parallel Sysplex is leveraged by z/OS subsystems such as JES2, CICS* Transaction Server, IMS* Transaction Manager, Db2* for z/OS and others.
The IBM Z architecture, as well as other system architectures, can also use GDPS to cluster across longer distances than within a data center. GDPS has several implementation patterns, from relatively simple storage replication up to support for full active/active CA. To learn more about GDPS, see “Raising the Bar on IBM Z Resiliency With GDPS”.
For clients running Linux* on IBM Z, the Linux OS and underlying hypervisors inherit HA features of the underlying hardware. Along with the hardware features, HA can be achieved through clustering of components such as Db2, Oracle or other software systems. They can also implement features of GDPS on Linux.
Tops in Resiliency
IBM Z clients use the platform for its security, scalability and its five nines of availability. CA has strong justification to protect businesses and prevent expensive outages. The combination of IBM Z hardware and software features makes Z resiliency tops in the industry, and helps businesses bounce back from outages and prevents them from ever happening.
Bill Seubert is the lead architect for IBM’s North America Z Technical Sales team.
Stephan Wiedemer is an offering manager for resiliency on IBM Z.
Sponsored Content3 Unknown Risks in Your Resiliency Armor
Post a Comment
Note: Comments are moderated and will not appear until approvedcomments powered by Disqus