Public cloud: Real-world examples of strategic success

March 25, 2020

Originally published on CIO

The public cloud has gone mainstream, with companies leveraging it as a strategic tool for digital transformation. IT leaders lend advice on migrating to public cloud services to drive business agility and innovation.

Public cloud services are a strategic weapon for CIOs. More than a way to shutter data centers and outsource computing resources, the public cloud offers CIOs the ability to focus on strategic projects aimed at boosting the bottom line.

Because of this, the public cloud has become the go-to platform for modernizing existing business applications and building new digital products, such as mobile software that delivers new customer experiences and machine learning systems that generate insights.

With most enterprises proclaiming “cloud-first” strategies, spending on IaaS (infrastructure-as-a-service) will reach $50 billion in 2020, up from $40 billion in 2019, and will grow 24 percent a year to top $74 billion in 2022, according to Gartner. “Cloud adoption is mainstream,” says Gartner analyst Sid Nag.

IT leaders shared with their experiences and lessons learned in making a strategic shift to the public cloud.

Finding room to grow in the cloud

Since 2014, Choice Hotels International has migrated more than 1,000 applications, including its ChoiceEdge core global reservation system and distribution platform, property management system and data analytics systems, to Amazon Web Services (AWS) for more than 6,900 hotels worldwide.

Choice is both modernizing legacy applications and building cloud-native apps atop AWS, Chris Judson, Choice vice president of engineering, tells Choice is also leveraging machine learning (ML) tools from AWS to support initiatives such as guest personalization and fraud detection and to reduce overbooking room inventory. “Our data is core to us and is one of the differentiators that lets us provide value to franchisees,” Judson says.

Along the way, the company has built a DevOps pipeline for running applications in AWS. Choice previously built a CICD (continuous integration continuous delivery) tool from scratch before adopting a suite from startup Harness, which features continuous verification capabilities, Judson says.

Judson’s advice: Migrating your infrastructure to the cloud doesn’t mean you can sleep on cybersecurity. The AWS-Capital One and Cloud Hopper incursions reinforced the point that CIOs must stay atop their cloud portfolio by ensuring every port and entry point is closed and that personnel are properly permissioned. “The cloud doesn’t remove the responsibility of knowing how to use the tools and understanding basic cyber-hygiene,” Judson says.

Data hygiene in the cloud

Easily accessing information to generate business insights is a top priority for health-care company McKesson, which shifted information stored in several legacy repositories onto Snowflake’s data warehouse software, running in Google Cloud Platform (GCP).

GCP’s ability to store sensitive data and scale capabilities in the cloud made it an obvious choice for running Snowflake, though the company also consumes compute services in AWS and Microsoft Azure, says Brian Dummann, chief data and analytics officer for the company, which distributes medical equipment and operates oncology clinics.

After re-engineering its data ingestion practices, McKesson began running Snowflake in production within 90 days of implementation. “I’m proud of the speed to execution,” says Dummann, adding that more than 3,500 business users are pulling down analytics from Snowflake. “It’s less about cost and more about speed and agility and the ability for data and analytics to create value in days or weeks rather than months.”

Dummann’s advice: When evaluating a cloud service, avoid vendor lock-in by picking a partner who will allow you to run applications that aren’t proprietary to that platform. “The trick is to choose partners who are cloud-agnostic so that you can run capabilities across more than one cloud,” Dummann says.

Oil and gas provider drills into the cloud

In 2016, BP joined the cadre of companies hitching their wagons to “cloud-first” strategies by building a multi-cloud architecture, says Stewart Fry, global vice president of enterprise IT services.

The company, which runs on AWS and SharePoint and other productivity and collaboration apps atop Azure, is shuttering data centers in Houston and London. Rather than lift-and-shift existing apps from its on-premises systems to the cloud, product teams “completely re-engineer” apps as cloud-native software, Fry says.

“We are getting away from justifying putting more compute into our physical data centers to exploit the opportunities the cloud is creating for BP,” Fry says.

Fry’s advice: Because making the most of the cloud requires several learning curves, it behooves IT leaders to lean on partners, Fry says. BP, for instance, leverages software from CloudReach to manage its cloud spend and utilization. “We realized pretty quickly that we needed specialist help to get things right from the outset,” Fry says.

Retail tech firm finds clarity in the cloud

Narvar, whose technology helps customers of Walmart, Gap, Adidas and other retailers track and manage online deliveries, began its cloud journey with AWS in 2015. Then politics happened.

As it turns out few traditional retailers want to support AWS’s quest for cloud domination. So, when Walmart complained about competing interests in Narvar’s use of AWS, the company switched to GCP, says CTO Ram Ravichandran.

Architecture preference was one of the reasons Narvar went with GCP over Azure. GCP supports Kubernetes and Knative for container orchestration and serverless computing, both of which Narvar required to run its business. These Google technologies enable Narvar to automate policies for computing and storage. “The technology was superior for our needs,” Ravichandran says of GCP.

To assist with the migration, Narvar implemented YugabyteDB, an open source distributed SQL database, which helps the company scale internet applications to onboard new customers during peak shopping season such as Black Friday and Cyber Monday.

Ravichandran’s advice: Moving from one public cloud vendor to another requires a lot of planning and architectural jujitsu. “The way to think about this is almost like moving houses: It is difficult but it can be done,” Ravichandran says. That is, you must continue to run the business with the appropriate capacity while ensuring that data is in the right containers — similar to furniture and other possessions — as you move between clouds.

Cloud fits clothing-as-a-service company

URBN, which operates the Urban Outfitters, Anthropologie and Free People brands, picked GCP from the outset when it entered the clothing rental market with Nuuly in 2019.

Nuuly runs its ecommerce site, warehouse management system, as well as a system powering its distribution center on GCP, using microservices and Kubernetes containers, says Chirag Dadia, Nuuly’s director of engineering.

Leveraging Google’s ML capabilities, Nuuly also funnels information about website events into a reporting dashboard to provide brand stakeholders real-time information about the business.

“Whether it’s our buying team evolving the product assortment based on this data, or our marketing team creating a targeted customer message, GCP allows us to offer our customers new ways to interact with the brand and make those optimizations nimbly,” Dadia says, adding that the Nuuly service launched within 10 months of its conception.

Dadia’s advice: Don’t skimp on planning your cloud architecture upfront. As the engineering leader, Dadia conducted plenty of initial analysis to land on the best cloud partner. “The overarching challenge we faced was creating the best solution for the job within a tight timeframe,” Dadia says.

Related Posts