I recently had the opportunity to share our journey with the attendees at DevOps World within the context of the Main Stage Keynote about what it takes to be an elite performer in DevOps. We’ve made some incredible strides and like all of us, there’s plenty more to do on the road to “continuous everything” and delivering perpetual value for customers.
Our Acceleration Journey
For us, elite performing is all about acceleration. To succeed, we must create simplicity amid our complex software development ecosystem. Our approach is really centered around the adoption of modern DevOps engineering practices, which is enabled by best-in-class developer platforms as well as the right tooling for every phase of the software development lifecycle (SDLC).
In order to simplify, it is also very important for us to have a set of patterns and standardization around a common set of tools, avoiding tool sprawl to ensure a high degree of predictability in our environment. These capabilities must have compliance checks, security scans and threat modeling built in from the start. Because, at the end of the day, our success boils down to having a developer platform that is highly intuitive, friction-less, easy to use and “always on.” And in many ways, it also needs to be invisible to our developers, so they can really focus on delivering value – instead of on tools – as code seamlessly travels across the phases of the software development lifecycle (SDLC), from development to test and to production.
Utilizing Machine Learning
Given the volume of builds, scanning and deployment data, machine learning (ML) is also helping us accelerate. With ML, there is a significant reduction in code duplication, making it easy for code reuse, code search and order generation of tests for repeatability, and all of these capabilities are in context within the integrated development environment (IDE) of a developer. With change-based testing, we are able to accelerate our deployment efficiently.
JET Centralized Developer Platform
We’ve brought all these capabilities together into a centralized developer platform called JET – JPMC Enterprise Toolchain—aimed at helping us achieve our performance and availability service level objectives (SLO). It is cloud agnostic and leverages industry leading components that are battle-tested. Designed for scale, the platform is built upon a continuously available architecture, instrumented with modern observability tools, with self-healing capabilities baked in and is heading toward a truly autonomous platform.
As our SDLC practices evolve and mature, we will continue to seek ways to improve the speed of delivery with increased predictability. In terms of outcomes, we’ve seen significant improvement in time to market, concept to delivery, as well as a greater frequency of releases. We’ve seen two-fold improvement in stability metrics with auto-remediation and advanced testing practices. And, we will continue to see significant improvements with the use of canary testing and feature toggles.
Now, shifting extreme left on our security and compliance focus, right from the design phase, we’ve seen significant reductions in critical security issues in production. We’ve been able to accelerate in our modernization journey with more workloads being deployed in hybrid and multi-cloud configurations.
It’s a team effort! We know one thing for sure, our work is having a meaningful impact on the value we provide to JPMC customers around the globe and that’s what drives us to continuously advance.