Packet Processing Primitives for Custom Data Plane Services
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network solutions struggle to provide highly customizable and efficient packet processing functions that cater to diverse and ever-changing customer demands, making it difficult to integrate these functions across scattered network elements while ensuring high performance and low latency.
Innovation Solution
A system and method utilizing APIs to create, customize, and configure data plane packet processing network solutions through a data plane service implementation (DPSI) platform, which includes a primitive creation module, application logic creation module, and service creation module, allowing users to build and manage network services via a service orchestrator and configurator.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If packet processing functions are implemented in scattered network elements, then existing network infrastructure can be utilized, but integration difficulty increases and performance cannot be ensured
Solution Approach 1:
The packet processing function is segmented into discrete network functions that can be independently implemented in scattered network elements (routers, switches, firewalls). Each network element performs specific packet processing tasks, allowing existing infrastructure to be utilized while maintaining modular architecture that simplifies integration.
Solution Approach 2:
A controller acts as an intermediary between the packet processing functions in scattered network elements and the central management system. The controller coordinates packet processing operations across multiple network elements, ensuring proper integration and data flow management without requiring direct complex interconnections between all network elements.
2Ease of manufacture
If packet processing functions are implemented in scattered network elements, then existing network infrastructure can be utilized, but performance (high-speed, low latency) cannot be ensured
Solution Approach 1:
Time-sensitive packet processing functions are extracted from the general packet processing pipeline and implemented as dedicated high-speed functions in network elements optimized for low-latency operations. This separation allows critical packet processing operations to bypass slower processing paths while utilizing existing infrastructure.
Solution Approach 2:
Different network elements are assigned different packet processing functions based on their local capabilities and optimization. Network elements with higher processing capability handle time-sensitive packet processing, while others perform simpler forwarding functions, ensuring overall system performance is maintained through localized optimization.
3Speed
If packet processing functions are integrated at deployment, then high performance can be ensured, but adaptability to diverse and ever-changing customer demands decreases
Solution Approach 1:
The packet processing function architecture is designed to be dynamic rather than static. Network functions can be dynamically configured, modified, and reconfigured based on changing customer demands while maintaining high performance through optimized data paths. This allows the system to adapt to diverse requirements without requiring complete redesign.
Solution Approach 2:
The controller and network elements are designed with universal capabilities to support multiple packet processing functions and customer requirements. The same infrastructure can serve different customers with different packet processing needs through software configuration, providing both high performance and adaptability.
4Adaptability or versatility
If all packet processing functions are built manually, then customization is possible, but complexity and time required for deployment increases
Solution Approach 1:
Common packet processing functions and network configurations are pre-configured and stored as templates or service profiles before deployment. When a customer requires packet processing services, the system retrieves and deploys appropriate pre-configured functions, significantly reducing deployment time and complexity while maintaining customization capabilities.
Solution Approach 2:
The system provides self-service capabilities where customers can configure and deploy packet processing functions through simplified interfaces without requiring manual technical intervention. Automation handles the complex tasks of function deployment, configuration, and integration, reducing deployment complexity while enabling customization.
Data Source
AI summary
A method and a system including a data plane service implementation (DPSI) platform operably accessible via application programming interfaces (APIs), implement packet processing network functions as a service. The APIs allow creation, customization, and configuration of network service-supporting application logics (ALs) and network service packages. The DPSI platform creates a packet processing primitive (PPP) by applying a use-case-specific primitive specification to a corresponding primitive template retrieved from a repository, and stores the PPP in a repository. Each PPP instance is implemented as an independent, self-contained, functional part of a whole data plane packet processing network function. The DPSI platform creates an AL based on a use-case-specific AL specification using one or more PPPs corresponding to the use-case-specific AL specification, and stores the AL in another repository. The DPSI platform creates and stores a network service package including one or more ALs corresponding to a use-case-specific service specification in another repository.