To Achieve Scalability, Don’t Overlook Your Mainframe
Scalability is the be-all, end-all of the IT world today. And when it comes to achieving scalability, mainframes can often help you more than commodity servers. Here’s why.
Scalability refers to the ability of IT infrastructure or software to support workloads of increasing size. Being able to scale is essential for ensuring that the extent of your IT resources doesn’t stifle your business’s ability to handle more transactions or support more customers.
Today, when most people think about how to achieve scalability, the go-to solution is the public cloud. On a public cloud, you can purchase a virtually unlimited amount of resources, ensuring endless scalability.
Yet if you own a mainframe, you may have another scalability solution already at your fingertips. While mainframe infrastructure is probably not the first type of IT resource that comes to mind when most people think about scalability, the fact is that mainframes can in many ways scale even faster and more cost-effectively than the cloud (or, for that matter, on-premise commodity servers).
Here are several ways in which mainframes can scale better than the cloud — and reminders of why mainframes remain valuable even in the age of the cloud.
Transactions Per Minute
Mainframes are beasts when it comes to processing transactions. While the exact number of transactions that a mainframe can handle each second will vary from one machine to the next, figures above one million are not unusual.
Again, that’s transactions per second; per day, you’re talking billions of transactions.
Can your cloud match this figure? Maybe, if you buy a huge number of virtual servers. But while the numbers will vary from case to case, there’s a pretty good chance that at the end of the day, your cloud computing budget will hit its ceiling before your mainframe runs out of transaction processing capacity.
Operating System Virtualization
One of the reasons people move workloads to the cloud is that they can create as many operating system instances as they want using virtual servers. Plus, they can use multiple types of operating systems at the same time.
That’s great. But what’s greater, at least for certain use cases, is that mainframes can also support effectively as many virtual operating system instances as you need.
What’s more, mainframes can host not only Linux instances, but also z/OS, all on the same hardware. Good luck finding a public cloud that lets you run z/OS.
What all of this means is that mainframes provide agility, which is the mother of scalability, so to speak.
Mainframe storage systems were easy to expand by adding new hardware long before software-defined file systems brought similar scale-out storage functionality to commodity servers. In this sense, mainframes have always been ahead of the scalability curve.
And while cloud-based storage is easy to expand by purchasing more storage space, it can also be expensive. Most public cloud providers charge not only a flat monthly or daily storage fee, but also impose extra fees for each time you access data or move it on the network.
With a mainframe, your storage costs are much more fixed. They’re tied pretty closely to storage acquisition costs, without much overhead on top of that.
In the public cloud or on commodity servers, accommodating a larger workload typically entails scaling “horizontally.” That means you add more servers to support more instances of the application that you want to scale up.
Mainframes could do the same thing using virtual operating systems. But mainframes also enable “vertical” scalability: Because of their immense processing and storage power, they can host a single application that accommodates a very large workload.
In many ways, vertical scalability is better than horizontal scalability. The more servers and application instances you add as you scale horizontally, the trickier it becomes to manage all of them and keep them secure. If you can instead scale vertically by supporting a huge workload with a single application instance, you only have one host and application to worry about.
This is like the difference between expanding your existing house to accommodate a growing family (that’s vertical scaling) and building an additional house, then spreading your family across your two houses (that’s horizontal scaling). Which approach seems easier to manage and more effective in the long run?
Download our eBook and discover how to address the top 5 mainframe security vulnerabilities.