FaaS Resource Allocation via Function Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In Function as a Service (FaaS) platforms, the limited size of serviced functions and the use of pre-generated common containers lead to inconsistent response speeds, resulting in delay times and resource wastage.
Innovation Solution
Implementing a system that allocates optimized resources based on function characteristics and execution information, using a prediction model to anticipate execution requests and allocate necessary resources in advance, such as containers and invokers, to ensure efficient execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a common container is generated in advance for higher response speed, then response speed is improved, but resource wastage increases and uniform response speed is not guaranteed
Solution Approach 1:
The patent segments the common container into function-specific containers based on execution information and characteristics. Instead of using a single common container for all functions, the system divides containers into dedicated units for different functions, allowing optimized resource allocation for each function while maintaining the benefits of pre-generated containers.
Solution Approach 2:
The patent applies local quality by allocating different resource configurations to different functions based on their specific characteristics and execution information. Each function receives a container with resources tailored to its needs, ensuring that resource allocation is optimized locally for each function rather than using a uniform approach.
2Loss of time
If a common container is generated in advance, then response time is reduced, but delay time is generated due to non-uniform response speed
Solution Approach 1:
The system segments the container allocation into function-specific units, ensuring that each function has a dedicated container with consistent performance characteristics. This segmentation eliminates the variability in response speeds that occurs when multiple functions share a common container.
Solution Approach 2:
The patent changes the parameters of container allocation by using execution information and function characteristics to determine specific resource configurations. This ensures that each function receives a container with parameters optimized for its requirements, leading to uniform and predictable response speeds.
3Ease of operation
If resources are allocated uniformly, then resource management is simplified, but resource efficiency decreases
Solution Approach 1:
The system implements self-service by automatically allocating resources based on execution information and function characteristics without requiring manual intervention. The FaaS manager autonomously determines the appropriate container configuration for each function, maintaining operational simplicity while achieving efficient resource utilization.
Solution Approach 2:
The patent dynamically changes resource allocation parameters based on execution information and function characteristics. This allows the system to optimize resource efficiency automatically while keeping the management process simple through automated decision-making based on observed patterns and function requirements.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A system for providing a function as a service (FaaS) is provided. The system includes a communicator which receives a request for setting resources to execute the function, a memory which stores one or more instructions, and a processor. The processor executes the stored instructions. When the processor executes the instructions, it analyzes characteristics of the function and provides recommendation information related to the setting of the resources to execute the function based on a result of the analyzing.