Redpoint’s Patrick Chase: Serverless Computing as the Next Evolution in Software Development
ABSTRACT
KEY POINTS FROM PATRICK CHASE'S POV
Why is serverless computing such an important category moving forward?
- Improved workflows for software engineers and data scientists. “I remember in the early days of Hadoop and Spark, we would kick off a distributed join and then go get coffee because it would take 15 minutes for a single job to run,” says Chase. “When we came back, we’d have to hunt for the logs in a sea of machines to find out what went wrong. My dream is that soon we’ll be able to run and debug code at any scale as easily as if it were running on my laptop.”What are the use cases attached to this category?
What are the use cases attached to this category?
- Serverless approaches have found a wedge into specific sectors with volatility in compute needs. The first use cases have addressed pain points associated with fluctuating compute demands, including ML training, streaming computation, and data transformation. “As of now, the biggest advantage to building on serverless systems, capable of quickly scaling up and down, has been in these use cases,” he says.
What are some of the potential roadblocks?
- A lack of interoperability between early stage serverless computing models and incumbent tech-stack tools are obstacles for adoption. “Serverless platforms sometimes require engineers to change how they structure code, debug functions, or monitor applications. This can lead to challenges in adoption because engineers need to educate themselves on new tools or re-architect their systems,” says Chase.
VISUAL: THE FUTURE IS SERVERLESS
We are entering the "Serverless Era," with products like AWS's Lamba making adoption feasible for cloud users. Also, according to Mordor Intelligence, the market is expected to register a healthy CAGR of over 23.17% during the 2021 to 2026 period.
IN THE INVESTOR’S OWN WORDS
Something I always wanted as an engineer was the ability to run code on the cloud as easily as if it were running on my machine. In the future, engineers will write code as if it were going to run locally, and the cloud will scale up and down in real-time to give them a seamless workflow with infinite scale.
MORE Q&A
Q: What do other market participants or observers misunderstand about serverless computing?
A: "Many people think of serverless computing as being synonymous with AWS Lambda. I think of it as a much broader trend where engineers are able to abstract away from the infrastructure they are using. Engineers don’t really need to know where their code is building, they don’t need to know where their database is running, and they don’t need to know how many machines they’re using to deploy an ML model. It just works."
STARTUPS MENTIONED IN THIS BRIEF
Acknowledgments
Editor's note: Please also see this memo by Tomasz Tunguz and Patrick Chase on Redpoint's investment in MotherDuck.