When a company like Siemens talks about APIs, you can be sure there are opportunities for teams to learn something valuable. Siemens has built an incredible reputation across industry, infrastructure, transport, and healthcare, fueled by technical innovation that’s making factories more efficient, healthcare more advanced, transportation more sustainable, and other purpose-driven pursuits.
So imagine our delight when we got to hear from Sven Legl, IT Architect & Innovation Manager at Siemens AG and Daniel Matos, member of the Siemens Automation DevOps team, at API Summit 2024, to hear about Siemens's approach to API management and the benefits they've seen with APIOps.
Siemens’s move from monolithic systems
The teams at Siemens believe in the power of tech to improve lives. As Legl points out, Siemens’s hands in infrastructure, transportation, and health care mean many of us interact with their products daily.
Siemens has a massive technology ecosystem to manage, and the goals for its technology are just as grand:
- For infrastructure, Siemens wants to reduce the energy used by 30%.
- For mobility, Siemens is working to enable more network capacity for public transportation.
- For health care, Siemens is taking on digital transformations to improve IT systems.
In Siemens’s Digital Industries (DI) area alone — where Legl and Matos work — there are around 70,000 employees. Reaching their goals and starting their journey toward effective API management started with one problem: monolithic systems.
Siemens previously managed things with big, monolithic systems, which Legl said cost time to market on top of other issues. The first step on Siemens’s API journey was to slice down and decouple these systems into smaller pieces.
Siemens uses Kong Konnect for API management in its architecture. For a deeper dive, check out Siemens's API Summit 2024 session.
Today, in place of legacy monolithic systems, Siemens has developed multiple microservices in their environment. The API management for this new architecture runs safely and efficiently with Kong Konnect.
"When we’re developing software applications, we need to do it rapidly and efficiently. With more and more microservices, we also need to ensure we do not integrate them manually," Legl said. "We started with Kong Konnect for today’s environment."
A stellar strategy for API management, scaling, automation, and more
What did Siemens need to do to transition from legacy, monolithic systems to one that could take on their aggressive goals today? Legl laid it out:
- Siemens Digital Industries (DI) is based on Kong Konnect today. “All that we’re using, we’re using Kong Konnect for it,” says Legl. “For vitals, for the manager, for the dev part which exposes our APIs. As a central platform for automation.”
- With Kong Konnect in place, Siemens sets up its own clusters for data sensitivity reasons, choosing Kubernetes on AWS for these. Each environment gets a production, test, and dev cluster, with production and test clusters in high availability mode and synced to one another.
But how does Siemens get the APIs from their businesses into this central platform approach?
Legl explained how the teams at Siemens are scaling in this federated architecture, with Kong helping during the CI/CD and deployment phases:
A look at how Siemens scales its API management, with Kong assisting during CI/CD and deployment phases.
- First, customer enablement happens with businesses who want to expose their APIs.
- Next, Siemens delivers an OpenAPI template that can be easily reused by customers. This way, they can do their own tests and linting.
- Afterward, integrations can be tested and merged with Siemens’s source control systems. After smoke tests to make sure APIs are in the state that’s expected, they can be deployed to the production data plane and dev portal as well as backup creation.
What’s in the secret sauce for Siemens’s APIOps? Matos, who handles Automation DevOps at Siemens, explained.
Siemens has developed APIOps processes that increase automation, meet customer needs, and minimize disruptions to their environments.
- The five steps at the top are run for each environment, including generating the Kong configuration from each OpenAPI spec, merging these into one file, uploading it to Kong Konnect, and later deploying pods using Helm and the Kong Helm chart.
- Matos explained that their approach to APIOps is seen at the bottom of the above visual. An initial linting phase where API specs are checked for errors or inconsistencies, things are rolled into dev and test environments. Tests, re-tests, and possible rollbacks are enabled before prod deployment and eventual release.
- Overall, Matos's team has found a way to keep multiple data planes up to date, keep product environments up to date with minimal disruption, and saved customers from waiting very long for changes to be deployed.
“These APIOps processes allow us to focus more on our customers’ needs and less on the technical details of our tasks," Matos said.
The tradeoffs and business results Siemens is seeing
In figuring out their APIOps approach, the team at Siemens is seeking to find the right balance between speed and consistency, Legl said. More speed means more and more customers served on their API management platform — but they also want to ensure consistency in applied governance rules.
Because of this, Siemens has set up a governance layer supported by a dedicated platform team. After the API team defines governance rules, guidelines, restrictions, and templates, they can hand these off to the team managing the governance layer, freeing them up for further development.
With guidelines and templates for all API interactions in one central place, dev teams get automated, direct feedback from this governance layer.
“That is quite cool, as it gives us the freedom of really scaling it up without scaling the platform team at the same time,” Legl said.
As for the business benefits of Siemens using these APIOps processes and Kong Konnect:
- Highly increased speed. Legl said speed has increased because the team can fully automate delivery pipelines end to end. Onboarding is quicker. Operations are simplified and standardized. And best of all, it’s highly scalable.
- Faster time to market and lower costs. These are directly connected, Legl said. Between the automation, standardization, and other efficiencies in this architecture, Siemens is reducing their time to market and lowering costs on their teams as a result. “That’s the reason we’re pushing APIOps from a technical as well as a business point of view,” he concluded.
Want more insights into how other organizations have unlocked next-gen API management? Explore more customer stories or watch Kong customer sessions from API Summit.
APIOps - The Key to API Excellence: Unleash APIs' Full Business Potential
