The primary driver of change is almost always cost-reduction—and that includes your decision about moving off the mainframe. Most company’s platform transition assessment narrowly focuses on software and MSUs / MIPs costs. However, those who have transitioned off the mainframe often find that their OPEX costs have more than doubled while their workloads have remained stable—and all this after putting in years of effort into transitioning platforms.
Let’s consider the real cost of ownership:
- 96% of customers paid more to “re-host” data in distributed environments
- For 38% of customers, costs on the distributed platform doubled in comparison to System z
Mainframe is clearly the “deal of the day”. And, there are ways to even further minimize the total cost of ownership (“TCO”) of your mainframe environment.
Enterprises like the State of Oregon Enterprise Technology Services (“ETS”) continues to grow their mainframe environment—and is quickly realizing the benefits of that choice. ETS was able to deliver 15 percent more services at 20 percent lower cost on the mainframe.
Scale with Your Business Needs
The mainframe continues to provide the best transactional performance—a single server accomplishes more than all the Facebook servers combined. According to Marc Staimer of Dragon Consulting:
“CICS handles more than 1.1 million transactions per second worldwide. That’s more than 95 billion transactions per day. To put that in perspective, Google searches average approximately 60,000 per second. Facebook likes average approximately 30,000 per second. Consider that a single IBM z System mainframe CICS can handle roughly as many transactions-up to 2.5 billion/day-as all of the Facebook servers combined.
Consider the Costs
Distributed environments are priced on a per-license basis. So, what does that mean for you? It may cost more to purchase applications and system management tools to run on multiple servers. Be cognizant of the terms and conditions. Contractual engagements with distributed software products can be defined on a per-user, per-server or even per-core basis.
How Complexity May Hinder Your Business
Running applications across distributed systems requires more code, more complex paths and in most cases, deep and wide knowledge of various programming languages. Updates must be validated in multiple environments before being rolled out across several more environments, all of which requires more computing power and staff time than having a centralized environment.
Prepare Your Enterprise for Anything
You might think losing a single distributed server is no big deal, which is true IF all the failover logic exists. With mainframe, the 5 9’s of availability continues to ring true with almost 99.999% up time. Cost-effective failover is core to the mainframe.
The Definite Guide to Platform Cost Comparisons
Understanding mainframe pricing is fundamental to evolving your IT environment. We often see customers calculate the cost of the mainframe by comparing IBM and other vendors’ software costs to a distributed environment running only a fraction of the workload. It is like comparing a grape to a watermelon.
If you want to conduct a true cost comparison in your platform transition assesessment, you’ll need to consider: scope, scale, performance, data center, staffing costs and more. Here’s where to start:
| Mainframe Cost | Distributed Cost |
| Procurement of mainframe for MSU | Procurement of distributed servers to run the same size workload |
| Data center footprint for the mainframe server | Data center floor space for distributed servers equivalent to mainframe computing capacity |
| Electricity to power the mainframe | Electricity to power distributed servers |
| Cooling the mainframe | Cooling the distributed servers |
| Maintaining the mainframe engineering team | Maintaining the distributed engineering team |
| Maintaining a centralized mainframe | Maintaining all the distributed servers |
| Lifecycle of the mainframe | Lifecycle of distributed servers |
| Platform migration | |
| Associated risks with the migration | |
| Disaster recovery on the mainframe | Disaster recovery with the distributed servers |
| Time to complete disaster recovery on the mainframe | Time to complete disaster recovery with the distributed servers |
Optimize Computing Costs and Plan for Enterprise Growth
So, what’s next? Minimizing your TCO for the mainframe.
Use our checklist below to get the most out of your IT environment:
- Leverage zIIP specialty processors to reduce the cost of running workloads on general purpose CPs
- Review and optimize LPAR CPC location, LPAR capping and LPAR weights
- Identify high MIPS consumption patterns and the underlying cause(s)
- Spread out MIPS consumption throughout the day
- Leverage monitors to identify processor delays
- Take advantage of IBM pricing deals to transfer new workloads to the mainframe
- Collocate workloads for higher performance and efficiency
- Take advantage of sub-capacity pricing
- Forecast MIPS growth by application, LPAR or Sysplex and then proactively plan where and when to run the workload
- Standardize CA software to optimize software per MIP changes
The mainframe’s strengths can be leveraged to move your enterprise into the future of digital. And it can do it at the right cost!