by Clint Boulton

GM, Synchrony bet big on private clouds

Feature
Sep 06, 2018
Cloud ComputingIT LeadershipIT Strategy

GM's Randall Mott and Synchrony Financial's Carol Juel detail how private clouds enable them to build and scale applications with speed and agility. No public cloud required.

cloud computing - data center - network servers
Credit: Thinkstock

CIOs worldwide are rapidly adopting public cloud services, jettisoning computing infrastructure in favor of CPUs, storage and data analytics rented via the internet. These services, offered by Amazon Web Services, Microsoft, Google and others, enable enterprises to upgrade applications at high velocity in service of digital transformations.

Yet some enterprises have boldly built private clouds that mimic the functionality and tools of public cloud platforms. They apply cloud traits, including automation, self-service, and tracking and monitoring of resources, to their internal data center to create a dedicated cloud.

General Motors and Synchrony Financial are two such outliers, venturing where many organizations have gone and failed. GM built a private cloud during the course of CIO Randall Mott’s move to insource the automaker’s IT. Synchrony built a private cloud that CIO Carol Juel says enables the bank to innovate in lockstep with its retail partners.

The private cloud’s murky past

These are bold bets, as the path to private cloud is paved with problems. Private cloud often requires IT to spend millions of dollars on software suites, which it then layers on top of repurposed infrastructure and management systems. But once private cloud systems are up and running, they often fail to keep pace with the technical changes required to align with business strategy, Forrester Research analyst Lauren Nelson wrote in a 2017 report.

Sometimes, they suffer from a lack of focus, as engineers are distracted by shiny new tech. Talent turnover can also wreak havoc on homespun clouds, Nelson adds. Forrester found that 86 percent of private cloud environments fall short of achieving the status of true cloud. “Building an environment focused on new workloads is hard, and optimizing existing systems lacks differentiation,” Nelson wrote. “As a result, private cloud struggles to gain attention amid competing priorities.”

No wonder why Gartner found the public cloud market grew 29.5 percent in 2017 to total $23.5 billion, buoyed by deployments at Walmart, Exxon and many other enterprises.

Accelerating app delivery

While many banks are exploring public cloud services, Synchrony, which provides private-label credit cards to retailers such as Gap, Walmart and Amazon.com, needed to quickly stand up its own infrastructure in the wake of its 2016 spinoff from General Electric. “We had to prove we could stand on our own two feet,” Juel says of the greenfield cloud experience.

Learning from adoption of software-as-a-service (SaaS) applications such as Workday and ServiceNow, Juel and her team built a private cloud using software from Pivotal Cloud Foundry, on top of which engineers build new, cloud-native applications. They use VMware vSphere to manage virtual machines. The company’s servers, routers and switches “run across our rails,” supporting an environment that features dynamic auto-scale configuration, large-scale in-memory data stores, real-time data streaming and massive parallel processing, among other functionality.  

Such features help Synchrony build and evolve its digital capabilities, including an API ecosystem that allows the bank to share data with its retail partners. The API platform has been instrumental in enabling Synchrony to build and make available to retailers chatbots and mobile shopping apps.

Machine learning capabilities baked into the cloud enable Synchrony to deliver insights that help partners personalize retail services. For example, the company has more than 3,500 individual data points, including balance, payments, merchandise purchased, for each consumer, on top of which it overlays 300 behavioral attributes, such as shopping frequency, to create a deep understanding of how consumers prefer to buy.

Juel hired several cloud experts to help address the learning curves involved in standing up a private cloud. Aping SaaS apps can’t check every box required for a soup-to-nuts private cloud infrastructure. Moreover, the push required Synchrony to move quickly from waterfall to agile development, which meant applications previously built as monoliths were difficult to upgrade.

Ultimately, Juel says Synchrony’s private cloud has helped the company accelerate application delivery, thereby improving revenue generation and overall customer experience for its retail partners. “Speed is the new IP in today’s economy,” says Juel, who won an award at CIO 100 Symposium for her private cloud deployment.

Private cloud as innovation vehicle

Mott meanwhile faced different challenges at GM. After years of dealing with outsourcers at Walmart, HP and Dell, Mott was sick of shelling out money to third parties for technology services a well-run IT team could deliver better. So upon joining GM in 2012, Mott began swapping more than 20,000 outsourced contract workers for an internal staff of 9,500, all while whittling down the automaker’s siloed application stack.

“You have to figure out how to take out costs, so you can reinvest” in new technologies, Mott told his peers earlier this month at the CIO 100 Symposium. He automated dozens of disparate processes and directed his staff to innovate, which he believed would help the company drive more benefits to the business.

One of the foundational fruits of his IT team’s labors was Galileo, a private cloud comprised of homegrown technologies as well as those from strategic partners. Mott says Galileo enables him to “automate things that matter to me,” including anything from SAP ERP functionality to mainframe tools. The system has enabled Mott to run GM’s business at a twentieth of the cost of the previous computing infrastructure operated and maintained by outsourcers. 

Galileo is the bedrock for more innovation technologies, Mott says, including Maxis, a self-service analytics platform that GM continually refines. Comprising proprietary technologies and open source tools such as Hadoop and Spark, Maxis enables GM employees, including analysts, data scientists and software engineers, to glean insights about the company’s business.

Such insights are critical in an era when digital disruptions, including autonomous vehicles and ride sharing, are altering market dynamics for the automaker. Galileo and Maxis help Mott and his team keep the business in lockstep with these changes.

Asked about public cloud during the CIO 100 event, Mott said: “I’m not going there. Jassy and Perot look a lot alike to me,” a reference to AWS CEO Andy Jassy and enterprise services provided by H. Ross Perot’s company Perot Systems, which Dell acquired in 2009 and later sold to NTT Data.

The takeaway

CIOs can build private clouds, but it involves the right tech stack and talent, and a commitment to continuously adapt. Can you say increased appetite for risk?

Forrester’s Nelson cautions enterprises against lulling themselves into “a false sense of simplicity.” Such hubris has led many tech teams astray. “Private cloud is hard,” Nelson says, summing up what she learned from interviews with thousands of enterprise technologists over the past 7 years.