February 2, 2018 Mainframe Agility

Why Your Mainframe Is More Modern than You Thought

The mainframe has long been the workhorse for many medium-to-large organizations. It toils behind the scenes performing the majority of the heavy (compute) lifting.

But even out of sight, like most technology, mainframes have been continuously updated, enhanced and improved. The most recent mainframe from IBM, the z14, is arguably the most advanced business computing system in the world1. In fact, not only is the z14 extremely advanced, it only occupies floor space of roughly five feet wide by six feet deep.

The problem is, no one sees mainframes, so many don’t understand just how modern they are. The fact that mainframes are locked away in the datacenter lets people’s imaginations run rampant about what they look like and what they are and, worst of all, leads people to assume mainframes are somehow frozen in time, that current mainframes are the same systems that were invented in the 1960s.

IBM System 360 from 1964

IBM System 360 from 1964

As Modern as a High-speed Train

If someone were speaking about a modern high-speed train and they showed you a picture of a steam locomotive from the 1800s, you would look at them like they’re crazy. Everyone knows what a modern train looks like, and we know that their capabilities are vastly superior to their ancient origins. Just because you don’t see a modern mainframe, don’t be fooled into thinking they are not completely state-of-the-art in every way.

Trains are, in fact, a very good analogy to mainframes. Trains are very efficient at moving a lot of people or freight from point A to point B.Sure, you could use thousands of automobiles to move all those people/freight, but then you would have to manage all those individual autos, including maintenance needs and drivers, manage traffic problems and park all those autos somewhere.

Clearly you cannot use trains to solve every transportation need, and cars are far superior for many applications. While I’m sure the auto manufacturers would prefer you to scrap the trains and buy cars, the smart thing is to use the correct vehicle for the correct application. It is especially ludicrous if you’ve already invested in the train and the tracks and have a transportation system that works.

It’s exactly the same with mainframes. Mainframes are far superior at handling the system of record with large volumes of transactions (think credit cards), while distributed servers can be better at system-of-engagement tasks (though they also create IT complexity).

Unparalleled Transactions

IBM z14

Photo Credit: Connie Zhou for IBM

Most large organizations have very efficient mainframe systems that work perfectly day in and day out. What is not as widely known is that while the mainframe is indeed excellent at performing complicated business transactions efficiently and securely and at a vast scale, it is also very good at any high-volume transaction.

These other transactions are sometimes called microservices, cloud services or REST APIs. In fact, the mainframe is so efficient that it is used by Walmart for a variety of “services” that used to be handled by distributed systems.

For Walmart, “It is all about the service that is faster, cheaper, and more flexible than traditional IT provided,” according to an IBM Redbooks publication. The lesson here is to capitalize on the advantages and capabilities of the mainframe by making it just another tool in the computing environment.

Speaking of a large number of transactions, the Internet of things is another perfect workload for mainframes; it is a device-to-device communication that can generate literally millions of transactions per hour. The mainframe is tailor-made for an application that uses IoT input.

The larger point here is not to be stuck with the notion that any platform or computing language is better or worse than any other. Use the right tool for the job. Be customer centric! Start focusing on ideas that help your customers and stop focusing on computing platforms and infrastructure. It’s more important, and much wiser, to quickly move forward with innovation than try to recreate something that isn’t broken.

1The z14 is based on a 10-core processor running at 5.2 GHz. The z14 system can have 240 Processing Unit cores, 32 TB of memory (RAIM) and is estimated to be able to execute over 140 Billion instructions per second. The z14 processor has a variety of hardware assisted cryptography features that focus on pervasive data encryption. The z14 supports up to 320 concurrent 16 Gbps fiber-optic I/O connections and a new I/O link called IBM zHyperLink Express that is designed to support up to 32 ports with link data rates of up to eight gigabytes per second (equivalent to 64 gigabits per second each). [SG24-8451-00 Redbooks IBM z14 Technical Guide]

2American Railroads needs an average of only one gallon of fuel per mile to move 480 tons of freight.