Industry executives and experts share their predictions for 2019. Read them in this 11th annual VMblog.com series exclusive.
Contributed by Owen Garrett, Head of Product, NGINX
2019: The Year of Serverless APIs
Almost exactly a year ago, we were discussing the rise of serverless computing after it received huge emphasis from AWS during their 2017 re:Invent conference, where the term was seemingly slapped onto every feature. Whether it’s servers, data storage, AI, machine learning, or other engines, serverless has cropped up just about everywhere in the last 12 months – cementing itself as the new Platform-as-a-Service (PaaS.)
But what’s next for serverless?
In 2019, the use of serverless APIs will more than double, albeit from a relatively small base. Amazon Lambda will dominate, and existing users of Azure and Google will adopt each of those providers’ ‘functions’ interfaces. As a result, on-premise implementations will only be adopted in limited situations.
Despite providers’ limited service level agreements for serverless platforms, enterprises will follow early adopters as they gain confidence. 2019 will be the year where many enterprises will learn through early projects if serverless is a) the long-term future for applications, b) to be used in a hybrid manner as part of a toolkit or c) just hype. In most cases, my money would be on option b, but organizations will also need to evaluate their reliance on hosted or on-premise solutions.
What are the challenges?
Before fully embracing this trend in the new year, enterprises will need to figure out where serverless can fit into their existing architecture. Currently, serverless requires a combination of optimism – the hope the function will trigger in time; faith – if it does not trigger this time, it should still run the next time; and indifference – if it doesn’t run, whatever.
Serverless functions run ‘blind’ and don’t store state, so debugging or recovering from faults is hard. Most of us have been burnt by the hidden costs of putting prototypes directly into production. That said, there are plenty of safe and appropriate applications for serverless, and they are a great way to quickly prototype complex interactions between services.
What practices can enterprise architects do to take advantage of this?
As organizations move APIs to various environments such as production, testing, or staging, they must ensure consistency for each environment to prevent misconfigurations. But efficiently handling client requests is only one aspect of a successful API (albeit a crucial one). Businesses need to monitor and troubleshoot performance to ensure customer satisfaction and analyze traffic to maximize business value. Serverless APIs maximize performance by reducing the average response time to serve an API call and minimizes the footprint and complexity.
For organizations beginning their serverless journeys in 2019, start with a pilot project to gain confidence in the tooling and cultivate best practices relating to the development, deployment, and management of serverless functions. During this process, document everything, so when this hidden, remote service starts to fail for whatever reason, you can identify and troubleshoot quickly. This will ultimately set your organization up for serverless success…just in time to tackle the next trend in cloud computing!
*Bonus serverless prediction*
In the rush to the bottom, there’s a 50 percent chance that at least one cloud provider will accidentally create a gold-rush by underpricing their serverless platform to the extent that it’s profitable for cryptocurrency mining. With reports showing the rise of crypto-jacking, this will have significant implications for companies who deploy serverless applications on platforms like AWS Lambda and Microsoft Azure, where an attacker could establish their own vast crypto-mining operation.
##
About the Author
Owen Garrett leads the product and go-to-market strategy for NGINX’s web acceleration and delivery technologies. Owen has over 15 years of experience in software engineering and product leadership at companies such as Riverbed. Today, Owen uses his technical and management expertise to optimize NGINX products and customer satisfaction.
