Middle age sees CA turn to AI, analytics, and mainframes to stay young

CA Technologies execs talk about how they see AI, blockchain, and automation in the future of the company.

CA Technologies is now over 40 years old. And like all technology companies its age, CA is looking to stay relevant and up to date in an ever-evolving landscape. To do that in the realm of software development, the company sees a lot of its future relying on AI, especially within the realms of analytics and automation.

Speaking at the company’s annual CA World conference in Las Vegas late last year, various CA execs such as CEO Mike Gregoire and global CTO Otto Berkes spoke about the future of this software development stalwart.


Jarvis, AI, and analytics

“10 years ago there used to be BI companies; Hyperion, Cognos etc,” says Mike Gregoire, CA CEO. “Fast forward a few years, all of them got bought, and one of the reasons is application developers had to build reporting into all of their applications. It became tablestakes.”

“Every application company started to differentiate on how good their reporting was and then after reporting came analytics. Now it's not just analytics, you're trying to predict what's going to happen in the future. Just about every company, in the same way that they've included reporting into their applications, they're going to include analytics and predictive analytics."

Described as a toolset that acts as the basis for all of the machine learning capabilities within CA, Jarvis is becoming the backbone of CA’s AI and analytics capabilities, and one it’s pushing across its entire product range.

“We're collecting a lot of data, it's always been a big strength of ours, and we want to correlate that and start to use machine learning and then predictive analytics,” says Duncan Bradford, VP, Presales & EMEA CTO. “Jarvis drives consistency and gives us a lot of options for sharing across the whole software factory.”

“The technology we've built is quite unique in terms of the architecture and capabilities, and the whole concept of elastic is there's APIs so they can feed any data into there and correlate that data. We've got Agile Ops, Applications, Infrastructure, Network, and that's all being fed into Jarvis so that we can perform machine learning against that correlated data.”

Born out of its internal startup accelerator, Jarvis debuted in the company’s Mainframe Operational Intelligence product in 2016 to help predict when potential issues might arise, it has since been integrated into CA’s Agile Operations, API Management and Security lines of business. And it might soon become available to customers.

“We developed Jarvis for our own use, but we're exploring the possibility that customer are having the same challenges with analytics and machine learning that we are and potentially providing it a standalone product,” says CA CTO Otto Berkes.


Mainframes and blockchains

Much of CA’s history is built upon the back of mainframes. The company still maintains a presence in the market, and believes nothing will change for some time. Gregoire says IBM’s new z14 mainframes are evidence that there’s “no shortage of innovation” going on in the space, CA’s new Jarvis-based Operational Intelligence offering has been its fastest-growing mainframe product, and CA's CEO sees plenty of potential in the old-but-reliable technology.

“When you look at microservices and containers, if it's done properly you won't know whether that's on a mainframe, Azure, AWS, or your own data center, whether it's running on which variant of Linux.

“The mainframe is just another compute Cloud source and to the extent that you can make that easy to use, and from a transaction cost perspective, viable, I think you’re going to start to see more workloads move to the mainframe or at least stay in the mainframe.”

The fact that mainframes have been around for so long, just sitting there processing large amounts of information, also means it’s a valuable resource.

“There's so much data on the mainframe what people find is with the cost of moving that data into a SQL or Hadoop cluster - where you lose all your security context in the process – that it’s best to leave it on the mainframe and then do your work from the mainframe pulling those specific datasets when you when you need them.”

Going forward, there’s potential for mainframes to be the home of AI and distributed ledger (aka Blockchain) workloads as well.

“The mainframe is an incredibly powerful and efficient transaction processor, and we view blockchain as a next-generation type of transaction processing capability,” says Otto Berkes, CA CTO.

“A lot of our exploration currently is around mainframe and using mainframes to enable blockchain-based applications such as smart contracts or providing data governance and being able to keep track of data in a secure way.”


Low code & automating software development

At CA World 2016, Gregoire was keen to push the concept that every company is now a technology company, and at the 2017 conference he suggested leading companies should optimize and automate their software development to create what CA calls a ‘Modern Software Factory’. Now that digital transformation means every company is software-driven and can make its own software, the processes for that should be as automated and repeatable as manufacturing lines in a normal factory.

But having every company under the sun draw from the same software development talent pool comes with its own problems.

“We are at that significant milestone where financial service companies, banks, insurance companies and the like are today employing more software development and IT engineers than technology companies are,” says Ayman Sayed, Chief Product Officer, CA.

“With every business becoming a software business, we are running out of engineers to hire and, at the current trajectory, we cannot graduate enough high-skilled workers in IT and software at a fast-enough rate. This leaves us with no option but to drive higher degrees of automation and add intelligence to IT and Lines of Business.”

In order to make up for the shortfall in talent by making application development as simple as possible and let more people create software, companies such as Salesforce have released various Low-Code platforms; where application development is done through largely visual, drag-and-drop methods and the actual coding is done behind the scenes.

“That you can get a great customer experience and extract some of the complexity of the of the Computer Science away from the business professional, that's a net positive because it helps get ideas into production quicker,” says Gregoire.

“There's got to be somebody and there's got to be a company that really understands what happens behind the scenes. There's a responsibility when you put technology into the marketplace that you really truly have to understand how it works, and have a three-hundred-and-sixty-degree view of that.”

“When we first started playing with virtual machines, everybody loved it until it broke and then you didn't know where your application, what physical machine your application was running on.”

For those that do still write code, companies are looking to bring at least some level of automation into the process. CA isn’t the only company that sees more automation coming to software development: GitHub CEO Chris Wanstrath last year said that “the future of coding is no coding at all”, and also released a tool for automatically flagging up potential coding problems.

“The efficiency of creating new software is going to be paramount for the future,” says Marco Comastri, CA’s EMEA General Manager.


Also read:

CA CTO: becoming a software company requires balance
Will life begin again at 40-plus for software veteran CA?
GitHub CEO: “The future of coding is no coding at all”

Low-Code: What it can and can't do for your company