The CloudExpo conference in New York is always a good take for developers, architects, and managers who want to understand where the technology of cloud computing is headed next. Serverless computing appears to be that destination.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
As session presenter Doug Vanderweide from the Linux Academy — as entertaining a speaker as you’ll ever run into at a technology conference — puts it, the first thing you need to know about serverless computing is that, yes, there are servers. They’re just not yours.
Let’s back up a step and note that today’s cloud computing boils down to microservices and containers. Each offers profound benefits, though neither is perfect.
Containers are hot and it is microservices that makes them great, Vanderweide says. Microservices break work into small steps with APIs to handle them. You can manage functionality independently, streamline development, and save time with reusable code. Microservices work best when running in small virtualized environments, namely, containers, which are quickly deployed, inexpensive to run, easily scaled and orchestrated, and offer version control.
But, beware, Vanderweide says. Containers exist in a cloud technology ecosystem that’s changing daily. They’re prone to sprawl, can suffer from broken dependencies, and they are at the mercy of networking woes. Serverless to the rescue.
As Vanderweide explains it, serverless computing is anonymous, generalized virtual machine instances that are managed by the cloud provider. They’re provisioned when needed and de-provisioned when you’re done. They’re billed based on executions and resource consumption, not at an hourly rate. With a focus on triggers, inputs, and outputs, along with high availability and superb scalability, serverless is a great match for microservices.
The base operating system (Linux or Windows) is a general configuration that supports multiple languages (Node.js, Python, .NET Core, Java, etc.). The key to this is the provider can quickly provision instances because they are all the same no matter the corporate user.
The real allure may be in the stellar TCO (total cost of ownership) that serverless delivers. When you look at VM vs. function-based pricing for 2 million executions per month, consuming 4 GB-seconds (4 GB of memory used for one second) per execution, the differences are clear. Vanderweide says that works out to $279.74 on AWS and $220.97 on Azure, but, in a serverless ecosystem, a paltry $121.80 for Azure Function and $129.86 for AWS Lambda. Pretty impressive stuff.
Vanderweide calls this “the long tail of serverless.” For the cloud provider, the sameness of configurations for everyone makes greatly reduces the expense of providing them. That means each new instance is, well, instantly profitable.
Compared with containers, similar function workloads cost less to run and you never pay for capacity that’s sitting idle. Beyond that, automation, abstraction, and cloud vendor services can eliminate DevOps tasks (and possibly DevOps payroll, too). Infrastructure costs drop, the systems development lifecycle is simpler, server management is no longer your problem, and deployments are faster.
It’s not perfect, of course. Serverless computing can suffer from laggy startups of cold code. And it’s an immature technology that may leave you wedded to a specific cloud platform provider, at least for now.
Vanderweide sums up the advantages of serverless computing with a quotation from Greg DeMichillie, head of developer platform and infrastructure at Adobe: “In five years, every modern business will have a substantial portion of their systems running in the cloud. But that’s only the first step.” DeMichillie goes on to say, “The next step comes when you free your developers from the tedious work of configuring and deploying even virtual cloud-based servers.”
What’s your take on serverless computing? Are you still trying to catch your breath with (and I hesitate to use this word) “traditional” cloud computing? Too much too soon? Or are you ready to get out in front of the next wave and carve out a new career path — again? Share your thoughts about serverless computing; we’d like to hear from you.