Let’s face it. Success begets success. The one thing that most investors agree on is that a proven track record of having executed and exited a business is a pretty good indicator that the person in question can do it again. And when the exit we’re talking about was for $3.7 billion, that wants to be a commensurate level of attention we pay. And so it is for Jyoti Bansal, the founder of application monitoring company AppDynamics that was sold to Cisco earlier this year for an eyebrow-raising $3.7 B. Bansal hasn’t had much of a break and has quickly moved onto his next thing.
The broader part of that next thing is BIG the none too modestly named Bansal Innovation Group, Bansal’s startup studio that was founded to focus on some of the big technology problems of the day. BIG is Bansal’s attempt to create the next cohort of billion dollar plus businesses. And first up is Harness, the company that Bansal is now CEO of. Harness aims to eliminate the pain around software update delivery by applying machine learning (of course) to the deployment mechanism. Harness will gain an understanding of an application’s baseline environment and initiate automatic rollbacks when irregular activity is detected.
Harness is coming out of stealth today and both launching its product and announcing a $20 million Series A funding round led by Menlo Ventures and BIG. Alongside Bansal, Rishi Singh, former DevOps platform architect at Apple, is CTO and co-founder.
And while he’s only had a few months post AppDynamics, Harness wouldn’t seem to be vaporware or just the promise of a real product. It’s a real live platform in production – Jobvite is using it and is reporting a 1,000% reduction in deployment times and Build.com has reportedly reduced detection and rollback time for production deployments from around 45 minutes to 10 minutes or less and also reduced the number of engineers needed to monitor new deployments to just one.
Why is this a problem?
First thing’s first. Why does Harness exist and why is the problem it’s trying to solve so important. It should come as no surprise to anyone steeped in the “software is eating the world” thinking that today most organizations are under unprecedented pressure to deliver new software and technology to users faster than ever, and with zero margin for error. But traditional approaches to software delivery – the world of Waterfall and centralized command and control – remain complicated, highly manual, and risk-prone. Which is where continuous delivery comes in, the approach towards software engineering that allows development teams to deliver their changes to end-users more rapidly.
Harness aims to democratize continuous delivery and offer it to all organizations – to do so it is applying unsupervised machine learning to the continuous delivery process, such that the platform understands an application’s baseline environment and can automagically roll things back in the event of unintended consequences. Ed Rose, director of software development at Build.com, explains:
The machine learning aspect of Harness give us the peace of mind and confidence to conduct multiple deployments on any given day. In a pre-Harness world we would spend maybe an hour with several of our senior engineers monitoring activity post-release. We can now reduce that to one individual who oversees Harness as needed, freeing up our engineering talent.
So, what does Harness do?
According to the company, the key product features that harness offers include:
- Pipeline Builder – allows teams to build and execute complete CD pipelines with serial or parallel workflows across their applications, services, and environments.
- Workflow Wizards – enables teams to rapidly build deployment workflows with out-of-the-box support for canary deployments, blue/green deployments using cloud technologies such as AWS EC2, AWS Lambda, Docker, and Kubernetes.
- Continuous Verification – applies unsupervised machine learning to automatically verify application deployments, detecting performance and quality regressions from tools such as AppDynamics, New Relic, Splunk, Elastic and Sumo Logic.
- Smart Rollbacks – automates rollback to the last working artifact version and run-time configuration with no required scripting or code.
- Continuous Security – offers full role-based access control capabilities with the ability to manage secrets, compliance rules and audit trails across deployments.
MyPOV
I’ll say it again, success begets success. Add to that the fact that Bansal was intimately involved with the world of application delivery, both as a software house and as a vendor, during his time at AppDynamics and you have a pretty good indication of where this will go. Clearly, Harness isn’t the only CD vendor around, but with a track record like this, you’d have to think they’ve got a good chance of going big.
