Faster Builds with Remote Execution
Software developers typically build source code on their local machines. Our remote execution service distributes such builds across a cluster of machines and remotely caches the results. This can speed up builds by a factor of ten and more.
Our Remote Execution Service in Action
The animation shows an example run: our service builds Bazel from scratch in 2 minutes and 23 seconds, about 15 minutes faster than a local build on a quad-core machine.
The animation shows an example run: our service builds Bazel from scratch in 2 minutes and 23 seconds, about 15 minutes faster than a local build on a quad-core machine.
We support all clients that implement the open-source Remote Execution API, including Bazel, BuildStream, Goma Server, Pants, and Recc.
To try out our service, you can access our our trial cluster. Alternatively, use our customer signup form if you would like to test our software on an individual cluster or on-prem.
You can use our managed remote execution service or self-manage and license our product with support. We offer deployments on-prem or in the cloud (such as GCP and AWS), other deployments on request.
Why Remote Execution?
Please accept YouTube cookies to play this video.
By accepting you will be accessing a service provided by a third party external to engflow.com.
Imagine you are pouring yourself a cup of your favorite tea and you keep getting interrupted.
This is what software developers have to deal with every single day.
A survey revealed: software developers wait 7.2 hours per week just for tests and builds to complete on their computer.
At an annual salary of 100,000 US dollars, that’s 17,000 dollars wasted per engineer, or a loss of 1.7 million for a company with a hundred engineers. Worldwide, this adds up to a whopping $458 billion US dollars per year spent just on waiting.
Research shows that these wait times lead to slower product releases, unhappy engineers and ultimately less successful companies.
Here’s our solution:
We at EngFlow – guided by Ulf, Helen and László – have over 20 years of combined experience working on build systems. With that deep insight, we developed a service that spreads out the work in the cloud to significantly reduce those wait times – and it gets faster the more engineers use it.
This does not just make software development faster and engineers happier, it also increases standardization, engineering mobility and transparency, and overall leads to better products.
EngFlow.com – delivering solutions that keep engineers in flow.
EngFlow’s Differentiators
Our remote execution service is enterprise-ready and production-grade:
Platforms
We offer both a fully managed solution on GCP or AWS, and a self-hosted solution that can run either in the cloud or on-prem. Our remote execution service runs on GCP, AWS, and Kubernetes.
GCP customers will soon be able to purchase our remote execution solution through GCP Marketplace.
Remote caching
Our remote execution service leverages remote caching to speed up builds. In principle, if configuring it to use zero remote workers, our product delivers pure remote caching.
SLAs
We provide an availability service-level agreement of 99.9% for the managed version, and have reliable customer support and flexible coverage options if you are self-hosting.
Monitoring
We provide easy integration with monitoring tools like Prometheus, Google Cloud Operations (formerly known as Stackdriver), and Zipkin.
Sandboxing
In the managed service, all actions run in a Docker container. Clients can configure the Docker image to match their needs.
Hermetic builds
Builds completed with our remote execution service are hermetic. This means you can easily reproduce results, for example a specific version you deployed.
Autoscaling
Scalability
Our remote execution software scales on the number of machines, as well as on the number of concurrently executed actions with a fully distributed scheduler frontend. It has been successfully run with up to 1000 executors, each of which has a configurable number of available CPU cores, memory, and other resources.
Remote Persistent workers
Our software supports persistent workers. When possible, compilers are not shut down after completing a task, but continue running to handle the next action. This can speed up builds significantly by saving time required for launching the compiler, avoiding re-parsing of input files and similar jobs. In particular, we support persistent workers for Android, Java and Scala. See the documentation for how to enable this feature.
Documentation
Our detailed, user-friendly documentation includes a comprehensive description of the system’s design, step-by-step instructions for deployment, a command-line interface reference, as well as best practices. It is available on docs.engflow.com.
Feature Roadmap
Key features on our roadmap include more performance improvements, MacOS and Windows support, and a browser-based Build UI.
Trusted Experts
EngFlow is not just your product partner but an expert advisor in optimizing developer productivity. We set the bar high for engineering talent and product quality. Building scalable, reliable infrastructure is hard. Our team, led by co-founder Ulf Adams, is uniquely equipped for this task. Ulf is a build systems industry thought leader and spent over 10 years perfecting his skills while creating and open-sourcing Bazel, as well as managing its internal equivalent at Google. EngFlow is the only company that has significant Bazel expertise outside of Google, which is reflected in the unmatched standard of EngFlow’s remote execution solution.
Connect today
Customer Sign-up
Learn more about our enterprise-ready, production-grade remote execution service and how we can help you get started. We offer a fully managed solution, or a license if you want to self-host.
Free Trial Cluster
See for yourself how you can optimize build times using our remote execution software: we have set up a free trial cluster to make it easy for you to try it out.