What do chocolate, skiing, watches, and mainframes have in common?

If you answered “Switzerland,” and specifically the state of Bern, then can you please send me some numbers for the next lottery drawing?

I’m not suggesting that the state of Bern is somehow the home of mainframes. But they do coexist with the more common IT variety of open systems in the care of Bedag Informatik, an IT firm whose sole shareholder is the State of Bern. As a result of the government’s stake in Bedag, and the fact that most of its customers are government and official entities, Bedag must store data on premises. In other words, public cloud is not allowed.

Bedag found a modern and innovative solution to this challenge. And that solution uses Model9 and NetApp® technologies. 

Bedag started as a mainframe shop back in the 1980s. By their own account, they are by now a “small” mainframe customer, with most of their workloads having shifted over time to open systems. I look at mainframes the way that I look at yodeling—something that used to be common, but now only a few people know how to do it. Which makes mainframes kind of expensive, even by Swiss standards. It makes sense, though. The fewer the mainframes that are in use out there, the higher the cost per unit is. These costs are then passed on to the customers, and the customers are not happy, in a respectful way.

I recently spoke with Martin Voegeli, the team lead for mainframe engineering at Bedag. I wish that the interview had taken place slope side while enjoying a hot chocolate, but, alas, it was on Zoom while drinking a cold cup of coffee. Martin has been in the mainframe world for a long time, and I appreciated his patience in explaining the ins and outs (recently, mostly outs) of mainframes.

Mainframe environments are typically some of the most siloed infrastructures. Data is stored in proprietary formats and is protected by expensive software and FICON-enabled tape libraries or virtual tape libraries (VTLs). These closed ecosystems tend to innovate less and to cost more. Remember CA (formerly Computer Associates), which at some point was the third-largest software company in the world? When Broadcom acquired CA, it became difficult and expensive to do renewals. That really says something, because CA itself was reputed to be hard-nosed about license prices.

To help break out of those silos and to reduce costs with an innovative approach, Martin turned to a company called Model9. With Model9’s industry-leading solutions in cloud data management for mainframes, Bedag can back up mainframe data to any Simple Storage Service (S3)–compatible object store.

Model9 has a suite of products that help customers securely free their data from mainframes. Backup and restore are just one such use case. Customers can also move archived data off the mainframe. But even more interesting, they can transform mainframe data—including IBM DB2 and Virtual Storage Access Method (VSAM), sequential datasets, and partitioned datasets—to standard open formats such as JSON and CSV.

Martin saw Model9 as a way to break up some of the mainframe monopoly, but he still needed an S3 target, and that target had to be on premises. When he consulted with his open systems colleagues at Bedag, he had two requirements: It had to be enterprise grade and it had to be fast. That is, at least as fast as the backup to the local VTL was. 

Unbeknownst to Martin, other Bedag customers were also asking for an object storage solution. The open systems team had their eyes on a NetApp StorageGRID^® object-based storage infrastructure, but they weren’t sure that they had enough demand for an entire multisite grid. Martin’s use case was serendipitous: He was going to be the anchor tenant that justified the initial two-site deployment. Other use cases (by now, 12 and counting) would become separate “junior” tenants that were capacity limited. The Bedag StorageGRID namespace spans two sites, approximately 100km (about 62 miles) apart. And Bedag uses StorageGRID information lifecycle management (ILM) policies to place a copy of each object on each site.

How significant, you ask? By moving away from outdated methods and by embracing innovation, Bedag has seen a cost savings of about 50% compared with backing up to VTL.

Unlike sequential VTL operations, object storage supports parallel requests, improving the overall backup throughput and reducing backup and restore times. Backup for Bedag is now significantly faster than the FICON direct-attached solution.

More Model9 and StorageGRID technology is a good bet. Bedag is testing Model9 Gravity, a product that transforms mainframe data into open systems format. Those datasets are stored on StorageGRID, and they’re made available to artificial intelligence and machine learning (AI/ML) and analytics applications that run on open systems through the S3 protocol. Model9 Shield can also take advantage of the StorageGRID S3 Object Lock functionality to protect backup data from ransomware.

The Model9 and StorageGRID architecture offers many advantages. For example, you can:

• Augment existing data lakes with your mainframe data.
• Reduce risk, because your mainframe data and applications aren’t modified.
• Lower your costs by avoiding expensive mainframe CPU consumption and by offloading those operations to object storage.
• Maintain copies of your mainframe data on a separate, air-gapped, and geo-distributed resilient infrastructure.

For an overview, check out the following Model9 and StorageGRID architecture diagram.

Model9 and NetApp technologies bring the benefits of open systems, standard protocols, and data portability to previously siloed mainframe environments. And you still get the same mission-critical, enterprise-class support that you’re accustomed to from your mainframe.

So, are you ready to lower your costs, reduce risk, and enhance access and protection for your mainframe data? To learn more about our integrated solution that has helped Bedag and other innovative customers solve multiple business challenges with a smarter and cost-efficient approach, read our joint solution brief.