For years, tech industry watchers have predicted a coming “cloud migration” that would eventually see entire enterprise organizations running on infinitely scalable remote servers. That migration is now well and truly underway: According to Flexera’s annual “State of Cloud” report, 92 percent of enterprises have a multi-cloud strategy and 80 percent use a hybrid cloud strategy. And though the advantages of cloud-based infrastructure are self-evident — from lower operating costs to remote access — widespread serverless adoption poses a number of new challenges for enterprises and the organizations that build the services they use.
Two types of services exemplify the wider change that has occurred as part of the cloud migration: enterprise resource planning software, now totally cloud-based; and security and backup services, which have adapted as data has moved from on-premise servers to the cloud. For example, software company Epicor is currently undergoing its own digital transformation, shifting its signature enterprise resource planning software to a cloud-native architecture. Meanwhile, cloud data backup company Rubrik is currently building a solution to secure information housed within Microsoft’s cloud-based 365 suite of products.
Both projects offer a glimpse into how the enterprise cloud industry is evolving — away from on-premise servers and toward a future where everything is AI-enabled — and so we checked in with technology leaders at Epicor and Rubrik to learn more about the challenges they’re encountering and the solutions they’ve deployed.
Martin leads software development teams at Epicor, which makes enterprise resource planning software as well as CRM and other SaaS or on-premises software offerings for business clients in the manufacturing, distribution, retail and services industries. Lawrence, who started his career in the military before moving on to lead R&D teams at a number of U.S. tech companies over the last two decades, is currently overseeing the digital transformation of Epicor’s enterprise resource planning offering. The work involves building and refactoring client-based software to modern cloud-native architected solutions, focused on browser-agnostic capability.
Tech stack: Lawrence’s teams use a .NET Stack, Azure Cloud, SQL and Angular on the front end and a C# back end on the server side.
What are some of the technical challenges your team has to overcome during the digital transformation of Epicor’s ERP service, and how do you plan to overcome them?
Our first core challenge is to ensure our choices of technologies enable our “intelligent ERP” endeavors. We must ensure that we have the right people for the job, as we were a heavily client-based architecture, and our team reflected this. Reaching our ambitions for our digital transformation required us to invest in training our teams on new modern technologies like Angular. This was an excellent opportunity for our team members, but it took a bit of time to get everyone sufficiently trained to accelerate in building the new software.
How is your tech team structured, and what are the advantages you attribute to it?
We are structured to meet the needs of a growing organization. We have a full scrum framework for core growth projects with new product development teams focusing on solutions that don’t currently exist in the market and new feature functionality; a sustaining engineering team utilizing Kanban to prioritize incoming issues; and platform teams focused on underlying supporting architectures. QA is a shared resource embedded in Scrum and Kanban teams to ensure consistency and testing, which is primarily focused on automation to ensure repeatability and quicker responses to issues. Automation is the lifeblood of any agile development team — we can never scale if we are focused on manual testing.
“As ERP customers become more integrated — manufacturers moving to distribution and distributors moving to manufacturing — they are looking for solutions that bridge this divide.”
What technological trends are currently affecting your industry, and how are you working to incorporate or account for them in your products?
In general, incorporating AI and ML into everything is the latest trend in technology. AI is an enabling function to extend workforces, with bots assisting in manufacturing or distribution and enabling organizations to engage with customers and suppliers quickly. In our industries, ML is helping organizations understand trends in data to help predict the next event. Whether it’s applied to inventory, distribution or manufacturing, data is integral in understanding the state of the business and impacts what customers make, move and sell. Cloud migration is likely the most pivotal technology move for enterprise applications; originally they were thought to be better suited for on-premise servers, but over the past five years the move to the cloud has increased.
What are some of the long-term challenges that lie ahead for Epicor?
Our long-term challenges are similar to other industries — how do we enable enterprise customers to adapt to the changing workforce? How do our applications enable organizations to utilize technology to bridge the divide between workforce shortages and business demands? As ERP customers become more integrated — manufacturers moving to distribution and distributors moving to manufacturing — they are looking for solutions that bridge this divide. Our ERP software is poised to bridge this demand for more access to integrated solutions. It’s not just what your ERP solution can do for customers. Still, how your ERP solution integrates within a customer’s ecosystem is unlikely to find a single solution in any enterprise, so integration is fundamental.
As the adoption of cloud-based enterprise software like Salesforce and Microsoft 365 accelerates, Silicon Valley cloud data management company Rubrik has grown to meet the increasing demand for its cloud backup technology. In fact, Senior Engineering Director Leela Tamma is currently building a backup solution within the company’s managed cloud platform for Microsoft 365. The goal is to build a scalable and secure backup system that protects valuable data “both in-transit and at rest.” Tamma started at Rubrik last year, after more than eight years leading engineering teams on a number of Microsoft’s most popular products.
Tech stack: The components comprising Rubrik’s managed service are deployed as Docker containers orchestrated by Kubernetes. Golang is the company’s language of choice, while Scala implementation of GraphQL is used to expose data from the services to the user interface. Scala is also used in the index and search pipeline. The UI is built using React and Typescript. Python is extensively used for E2E tests and stress tests.
What are some of the technical challenges your team has to overcome while working on the Microsoft 365 backup technology, and how do you plan to overcome them?
At its core, cloud data protection requires taking snapshots of data at scale with an ability to recover some or all of the data as fast as possible — all the while operating within the constraints that are set by public cloud providers and enterprise cloud offerings such as Microsoft 365. In order to achieve high throughput and low cost of ownership and deliver data protection at scale, our engineering team is innovating to build solutions in a cloud-agnostic fashion in the data storage space and compute aspects.
For storage, we are working on how to efficiently store data through data deduplication, immutability, IOPS reduction, solutions to deal with failure as the primary mode of operation and garbage collection to meet various compliance requirements. On the compute side, we are working on supporting multi-tenancy, dynamic resource manager, auto-scaler, indexing/search functionality and are scaling out techniques to work within the constraints of enterprise SaaS offerings.
How is your tech organization structured, and what are the advantages you attribute to it?
Our tech organization is structured into data protection and infrastructure teams. The data protection team’s primary focus is to build the data plane to orchestrate data protection leveraging the APIs exposed by enterprise SaaS services. The infrastructure team focuses on provisioning and orchestration of compute and storage resources required by the data protection team, geared towards providing high availability and data recovery. This helps each of the teams build expertise, innovate in their primary focus areas and offer a best-in-class experience for customers.
What technological trends are currently affecting your industry, and how are you working to incorporate or account for them in your own products?
The pandemic pushed the majority of industries to remote work, thereby accelerating the usage of cloud enterprise SaaS offerings at an unprecedented scale. Along with scale challenges, we encountered challenges associated with security and increased attack vectors. This provides unique opportunities and challenges for Rubrik to expand its protection to include generic data protection that is cloud- and data-agnostic. It also made us think about how to build generic data protection that generates usage insights while simultaneously keeping core principles of product simplicity, compliance and ease of use.
What are some of the long-term challenges that lie ahead for your company?
Public cloud and SaaS adoption is growing so rapidly due to the agility that these environments provide. As a result, traditional IT practices like backups can impact agility. A problem that our industry will have to solve is to enable backups of such environments without impacting agility. As we tackle this cloud growth, a potential challenge that could lie ahead is scaling our product and people organization.