Before the COVID-19 pandemic, enterprises were at various stages of their cloud strategies, whether that meant finally moving their email server to the cloud, switching to Microsoft 365, or even aggressively exiting their own data centers and going fully cloud-native.
When the pandemic hit, it pushed all those plans into warp-speed. There was no time for extensive user acceptance testing, long rounds of strategy meetings, RFPs, and proofs of concept when enterprises had been told to send all their employees home almost overnight, or in the case of India, literally overnight.
In interviews with analysts and companies moving to the cloud, InfoWorld identified the three stages of modernization that enterprises are likely to follow.
Why the pandemic forced faster modernization
Worldwide enterprise spending on cloud infrastructure bypassed spending on data center hardware and software for the first time in 2020, as the COVID-19 pandemic accelerated a decade-long industry trend towards running enterprise workloads in the cloud, instead of on-premises servers or via managed providers, according to research by Synergy Research Group.
Red Hat CEO Paul Cormier, in a recent presentation, observed that the pandemic “probably accelerated things by five years” among its customer base. “Customers accelerated what they were talking about doing, like looking at their architecture as a whole, taking more to the cloud, adding providers, moving more apps to containers—even on-premises—and now they are using that opportunity to do it,” he said.
For many of these organizations, the pandemic simply put into focus the limitations of what we often call “legacy” technology—systems based on old, sometimes obsolete methods of computing for which there are better alternatives available today, but that are often expensive to replace. The pandemic changed the calculation, driving faster replacement of the legacy.
This leap in legacy modernization efforts can be seen across every industry, including shipping giants like Maersk committing to shifting its ERP systems into the cloud, retailers doubling down on flexible, distributed e-commerce systems, , and the British National Health Service moving all of its email systems online.
For the British retail group Sainsbury’s, the pandemic proved out the benefits of moving its e-commerce systems to the cloud, where they could be better flexed to meet volatile demand. “Without that, I think we would have really struggled to pivot and adapt the business as fast as we have,” Phil Jordan, group CIO at Sainsbury’s, said during the recent CIO Future of Cloud and Digital Infrastructure Summit.
Modernization stage 1: Enable remote work
Enabling employees to work from anywhere without having to pipe into legacy on-premises email, collaboration, HR, or document storage systems—essentially anything that could easily be switched to a software-as-a-service (SaaS) option—was the key first step for many CIOs in the spring of 2020, as the pandemic took grip of the world and forced many people to set up their home offices for the first time.
As a result, videoconferencing software like Zoom added more new users in the first two months of 2020 than it had in all of 2019, and by April 2021 it was hosting 300 million daily meeting participants. Use of rival collaboration platforms also surged, as Microsoft Teams had 115 million daily active users in the first quarter of 2021, Microsoft reported, and Google said it had 100 million participants logging into Google Meet meetings every day at the end of 2020.
“There were a number of impacts of the pandemic on our cloud strategy. Obviously, deployment of collaboration tools, video conferencing, and live chat was hugely accelerated and only made possible by the SaaS tools being ready to use and capacity being available,” Ian Haynes, CTO for global cloud services at bank HSBC, told InfoWorld via email.
Similarly, shipping giant Maersk had to quickly enable 40,000 employees to work from home, from what would normally be 10% capacity to 100% in just six weeks. Also, on the business side, Maersk saw a third of its smaller customers quickly switch to digital channels via the Maersk.com portal to purchase logistics services.
“It has definitely helped to accelerate the uptake of those digital products, as well as helping us internally in terms of how we collaborate,” Will Wigmore, head of enterprise architecture at Maersk, said during the CIO Future of Cloud and Digital Infrastructure Summit.
Modernization stage 2: Seize the cloud-native opportunity
Once employees are set up to work productively from home, many IT leaders will want to maintain their momentum by pushing more and more legacy workloads into a consistent cloud consumption model. Unlike one-size-fits-all SaaS solutions, many of these business applications bring with them a tougher set of decisions for architects and developer leads to make around what to lift and shift and what to rearchitect for the cloud.
“Generally speaking, the pendulum is swinging toward an ‘improve then move’ type of approach in which customers transform to some degree during migration,” said Forrester Research analyst Bill Martorelli. “In practice, customers tend to take a portfolio approach, encompassing a variety of approaches, including lift and shift, modernization, replace with SaaS, and retirement, depending on the workloads.”
For example, Nadine Thomson, global CTO at media agency Mediacom, says her engineers are already starting to “revisit applications which we lifted and shifted into the cloud” so that they can start thinking about how to containerize and optimize so that they are fit for purpose for years to come.
The pandemic created a particularly compelling modernization imperative for banks, which had to enable all major services to be accessible online. “If we needed any more incentive or demand for the cloud, the pandemic provided it,” said HSBC’s Haynes during the AWS Financial Services Cloud Symposium earlier this year. “We saw huge increases in online banking services, hundreds of government aid and benefit schemes to be implemented, and some very dynamic markets. This accelerated our deployments to the cloud. And, for existing workloads, we were able to react to some very dynamic demand.”
For example, the bank took the opportunity to rearchitect its equities risk calculations and mortgage-brokering applications using a “move and improve” approach. This meant lifting and shifting the application, while also “provisioning and maintaining infrastructure as code, using software defined networking and virtual services for load balancing and firewalls, and replacing middleware and database software with PaaS [platform as a service] services where possible,” Haynes said. “Then, once running successfully on a cloud platform, we use subsequent phases to refactor the applications into microservices and APIs, and introduce technology such as containers, serverless computing, or additional PaaS services.”
Fellow global bank Morgan Stanley has been focusing on its most compute- and data-intensive workloads when it comes to legacy modernization: its risk models. In particular, the equities risk model needs to crunch through at least three billion data points on a nightly basis, sometimes more, with the calculation running at the end of the day across a variety of geographies.
“Doing this on-premises means lots of spare capacity that sits idle,” Vikas Chawla, executive director at Morgan Stanley, said during the AWS Financial Services Cloud Symposium earlier this year.
“Architecturally, when we designed this workload, we intended it to be cloud-native,” he said, but it still required adoption of modern principles like infrastructure as code, deploying via Docker containers, and reading from object storage, all running on more economical AWS EC2 spot instances to keep costs in check. This gave Chawla and his team a blueprint to apply to other risk-based calculations at the bank, such as counterparts, credit, and derivative risk. Now, they are looking to take what they have learned and to apply it to workloads that might be less well-suited as they exist today to the cloud.
“These large workloads are one part of a broader strategy,” he said. “In addition to elasticity, we migrate workloads that benefit from taking advantage of cloud-only products. Examples include complex data platforms as well as when we are going through major application renovations or have brand-new products. …Successful use cases like this are important to drive our strategic efforts forward and ensure the commitment from the business.”
While banks like HSBC and Morgan Stanley are naturally taking a more cautious, centralized approach to rolling out cloud platforms to their developer teams, other organizations are looking to make more drastic moves by fully rearchitecting their core business applications to run cloud-native, as the fitness company Peloton did throughout 2020.
Like offices and shops, gyms and workout studios were also closed for long stretches of 2020 and early 2021, leaving many of their customers to either lace up their running shoes or invest in expensive home exercise equipment to stay fit during the pandemic. One big beneficiary of this shift was Peloton, which quickly packaged all of its workloads into containers orchestrated by Kubernetes so it could better scale up to meet rapidly increased demand for its popular virtual cycling and fitness classes.
“We scaled in nearly every dimension, sometimes in triple digits for app downloads, subscriptions, streaming, compute levels—every aspect of speed and delivery—and we did this quickly, while being 100% remote,” said Jim Haughwout, vice president of platform at Peloton, at the Kubecon Europe conference in May 2021.
“Cloud-native has been the scaffold of the COVID era,” said Priyanka Sharma, general manager at the Cloud Native Computing Foundation (CNCF). “The trade-offs are generally a balance between speed, expense, and continuity with functionality, performance, cost benefits, and disruption,” said Gartner analyst Ed Anderson. “Organizations with a low tolerance for risk, limited funds to invest in an expensive cloud migration, or that see their migration as a first step toward something bigger in the future are more likely to opt for a simple migration: lift and shift or lift and optimize. Those with funding and a tolerance for risk and disruption are likely to take the more progressive route.”
Modernization stage 3: Move the deep legacy apps to the cloud
The final piece of the puzzle for long-established enterprises is moving the deepest legacy workloads into the cloud. We are talking about 20- to 30-year-old supply chain or payments systems residing on mainframes and written in languages like Cobol.
“When you are thinking about your competitive edge, if you are on mainframe, you will fall behind. If you are going through a digital transformation, you will eventually have to deal with the elephant in the room of that big mainframe in the corner that is hosting 70% of your business applications,” said Tim Jones, managing director of application modernization at software service provider Advanced.
“We’ve got some big workloads, some very traditional mainframe-based workloads that we’d really like to have cross the Rubicon and get those in the cloud,” said Sainsbury’s Jordan. “We’re now into the real heavy lifting. There are some workloads like supply chain that retailers don’t like to modernize, because it’s the absolute core of our business. But we’re on with that now and making that a cloud-based service, with all the AI and the machine learning opportunity that comes off the back of that.”
Moving these workloads to the cloud is not easy, but it can be done. Take the UK Department for Work and Pensions (DWP), which is responsible for various welfare, pension, and child maintenance payment schemes that serve as many as 20 million claimants a year. Starting as far back as 2015 and only completed in January 2021, the agency opted for a conservative “like to better” conversion and migration of its Cobol applications to the object-oriented Micro Focus Cobol, hosted on private cloud servers by Crown Hosting Data Centers, a joint venture between the UK Cabinet Office and Ark Data Centers.
This included the migration of the DWP’s largest service, the Jobseekers allowance, over Easter 2020, just as the COVID-19 pandemic was starting to grip the country. “That was an intense time to do an application migration for a benefits service that was starting to see an avalanche of claims because of how COVID was hitting the country,” Mark Bell, the virtual machine environment replacement (VME-R) program lead at the DWP, told InfoWorld.
The new cloud-based Micro Focus Cobol setup may seem like a small step in the right direction, but it allows the whole organization to be more responsive and modern in its approach to software, even opening the door to a more drastic rewrite into something like Java or C# in the future, he said.