Bulk Up on Big Data from Your Mainframe, Part 1
“Big Data? We’ve spent years right-sizing our mainframe data center, our apps, and our staff. We’ve streamlined our streamlines. Now you’re saying we have unused data that we should be leveraging for some kind of big data warehouse?”
Bulking up on enterprise data — big data, with big iron — usually starts with supplements from offline data and moves out from there. But let’s first review how IBM is positioning the z Series to engage in a cloud-centric world where x86 is often the default choice.
From Archive to Live
IBM’s mainframe announcements last year demonstrated Big Blue’s commitment to delivering more horsepower per buck in flexible configurations to position IBM customers for hybrid cloud scenarios.
As Information Week’s Doug Henschen explained then, “Where big companies already have big-iron workloads, the System z has already proven its appeal in workload consolidation, private cloud, and hybrid cloud scenarios.” But he also said that mainframers still have to come to terms with the market as-is. He called it a “Big Data Reality Check.”
This yea, the same publication reported IBM’s announcement of the “all new” z13s mainframe. An entry-level box, the z13s increases the breadth of the mainframe segment by appealing more to midsize enterprises. IBM’s calculus is to include a hybrid cloud mainframe that natively implements encryption and decryption in hardware — doubling effective throughput. They’re also bundling tamper-resistant hardware, a beta version of Cyber Security Analytics, Security QRadar security, multi-factor authentication for z/OS.
With companies like predictive analytics specialist Zementis and MongoDB for databases announcing support for z Systems in 2015, many recognizable x86 Big Data tools have become available on the mainframe.
Companies in the space are confident that z Systems can play a role in the world of cloud. In fact, some argue that z must be at the center of any cloud strategy for any large scale enterprise. And they report that is is happening now, not in some hypothetical future.
For adherents of circular trends, the mainframe incarnation of cloud services returns to the place where it started in the 1970s. What’s new is Big Data Volume, Velocity and Variety; e.g., as shown by ZPSaver from Syncsort, by leveraging advanced analytics platforms and by new capabilities for Spark clusters to ingest mainframe/non-mainframe data sources, and similarly by the ability of mainframe programs to consume data from Spark clusters.
Analytics will get only more complex. It will be asked to perform more computations and do it in real time. The idea of “reports” will itself be transformed into something more nuanced.
Bulk Up Big Data Repositories — with Mainframe Supplements (credit: Yasunobu Hiraoka | Flickr)
Case Study: Radixx International and the Hybridized Mainframe
Orlando-based Radixx International has grown as commercial airlines sought new ways to offset downward pressure on fares. Radixx offers a full-service e-commerce sales and distribution capability to more than forty airlines worldwide. To achieve this, Radixx stores all critical apps on a z System mainframe and hosts front-end applications on IBM’s SoftLayer cloud. The approach gives the vendor speed and security for core processes running on the mainframe segment that, simply put, must not fail.
The need for different network architectures is also driving change. For example, a global automaker with more than 50 vehicle assembly operations wanted to standardize its factory communications. As manufacturing processes scale up to handle bidirectional data streams between mainframe manufacturing execution software and factory PLCs, robots, conveyors, welding, fill fluids and external devices, the speed and stability of the mainframe remains as crucial as ever. Firms like PTC, which recently acquired Kepware, whose products have served to bridge mainframes with factory process-control systems, are investing in these and related technologies to position themselves for scalable, data-rich environments.
Niche software, such as IoT PLM, Retail PLM, and SLM are examples of new application domains where speed of transaction flow, security, and real-time analytics will play an increasingly important role. As Big Data and IoT take hold, each industry will likely develop specialized appetites for data.
In Part 2, next Friday, I’ll talk about more use cases.