Function workloads
A Function is a stateless code snippet that performs tasks in response to an HTTP request. With IBM Code Engine Functions, you can run your business logic in a scalable and serverless way. IBM Code Engine Functions provide an optimized runtime environment to support low latency and rapid scale-out scenarios. Your Function code can be written in a managed runtime that includes specific Node.js or Python versions.
A code bundle is a collection of files that represents your function code. This code bundle is injected into the runtime container. Your code bundle is created by Code Engine and is stored in container registry or inline with the function. A code bundle is not a Open Container Initiative (OCI) standard container image.
Lifecycle of a function instance
When a Function is invoked, the corresponding Function instance is initialized with the configured Runtime container and Resource parameters. The process of the first initialization is referred to as cold start.
To reduce the cold start latency, Code Engine optimizes the invocation by pre-warming certain runtimes with specific CPU and memory configurations. Pre-warmed combinations for functions include Node.js 18 and Python 3.11 versions as well as the default CPU and memory combination for Functions, which is 0.25 vCPU x 1 GB of memory. In addition, the system is designed to improve the reuse of Function instances that are already initialized. Therefore, a Function instance is kept alive after the invocation is finished to allow subsequent invocations by reusing the same instance and reusing the state of the instance when the last invocation completed. The reuse of a Function instance is not guaranteed.
How do functions compare to apps and jobs?
| Characteristic | App | Job | Function |
|---|---|---|---|
| Execution time (duration) | Long-running (10 minutes per request) | Long-running (up to 24 hours) | Short-running (2 minutes or less) |
| Startup latency | Medium | Scheduled start | Low |
| Termination | Run-continuously | Run-to-completion | Run-to-completion |
| Invocation | On request or permanently running | Scheduled | On request, instant |
| Programming Model | Container-based build and execution | Container-based build and execution | Language-specific source code files and dependency metadata |
| Parallelism | Parallel execution, flexible | Low to medium parallel execution | High parallel execution |
| Scale-out | Based on number of requests | Based on job workload definition | Based on events or direct invocations |
| Optimized for | Long running, highly complex workload and on-demand scale-out | Scheduled or planned workloads with high resource demands | Startup time and rapid scale-out |
For more information, see Planning for Code Engine.
What are key features of working with functions?
Review the following topics to learn more about working with IBM Code Engine Functions.
- Isolation
- Logging
- Runtimes
- Running functions
- Security
- Invocation concurrency and scaling of function instances
- Packaging your source code for a function
Isolation
Code Engine is a multi-tenant, regional service where tenants share the same network and compute infrastructure. In particular, the network and compute infrastructure are shared resources and some management components are common to all tenants. However, tenants and their workloads are isolated from each other by using Code Engine projects. Code Engine prevents communication between projects, providing isolation to your applications inside a multi-tenant environment, which can include applications, batch jobs, and functions. In addition, there are access controls that are performed on a resource level to allow only authorized users to perform certain operations on project resources, such as Functions or other Code Engine workloads.
For more information about workload isolation, see Code Engine workload isolation.
Logging
Function code can write stdout and stderr logs that are captured and forwarded to an IBM Cloud Log Analysis instance. IBM Cloud Log Analysis is a Platform logging instances that you can set up to index your logs.
For more information, see Viewing logs.
Runtimes
Code Engine includes Managed runtimes that you can use for your Functions.
Managed runtimes include Node.js and Python versions and specific CPU and memory combinations. These runtimes are optimized for fast startup. These runtimes are pre-warmed, which avoids the overhead of pulling container images and starting containers and processes. Your code is injected into an already running container.
For more information, see Runtimes.
Running functions
Whether your code exists as source in a local file or in a Git repository, or your code is a container image that exists in a public or private registry, Code Engine provides you a streamlined way to run your code as an app.
You can create and invoke your Function in Code Engine in the following ways:
-
Run from an existing code bundle.
-
Run from existing source code. If you are starting with source code that is located in a Git repository or on your local workstation, you can point to the location of your source and Code Engine takes care of building the image for you.
For more information about creating and invoking Functions, see Working with functions in Code Engine.
Security
Code Engine provides immediate DDOS protection for your function. Code Engine's DDOS protection is provided by Cloud Internet Services (CIS) at no additional cost to you. DDoS protection covers System Interconnection (OSI) Layer 3 and Layer 4 (TCP/IP) protocol attacks, but not Layer 7 (HTTP) attacks. See DDoS protection.
Code Engine also provides a service mesh to use its networking layer, which enables mutual Transport Layer Security (TLS) traffic for functions, thus securing service-to-service and user-to-service communication.
For more information about security, see Code Engine and security.
Invocation concurrency and scaling of function instances
When multiple functions are being invoked at the same time, Code Engine initializes new function instances for each invocation, but at the same time, tries to maximize the reuse. Only one function invocation is handled by a function instance
at a single point in time. For Node.js, the function can be configured with concurrency > 0 to allow multiple invocations to be handled in a single function instance.
Packaging your source code for a function
Functions can be packaged in three different ways.
- as a single file
- as multiple files (with a folder structure and dependent modules)
- as a container image
How can I get started with functions?
To deploy a simple Code Engine application with a hello-world sample image, see Running IBM Code Engine Functions tutorial.
To dive deeper into working with functions, see Working with functions in Code Engine.