In the field of computer networking, traditional configuration methods have relied on humans using command-line interfaces. However, in the past decade, various network abstractions have emerged as alternatives to manual configuration. Despite their availability, there is currently no objective way to compare these abstractions and determine their effectiveness. As a result, there is no consensus within the industry regarding the direction modern network abstractions should take, leading to a lag in their adoption. This framework utilizes metrics derived from graph structures to evaluate the simplicity, efficiency, and effectiveness of different network abstraction models. By applying this framework, Moreno finds that while some existing network abstractions such as Kubernetes or Cisco Application Centric Infrastructure models are efficient for storing network policies, others like public cloud-based abstractions remain infrastructure-centric and suffer from excessive complexity. The paper emphasizes the need for standardized evaluation methods to objectively assess different network abstraction models. By providing a comprehensive comparison framework based on graph structure metrics, Moreno's research contributes towards addressing this gap in understanding and facilitating informed decision-making regarding modern network abstractions. Overall, this study highlights the importance of objectively evaluating network abstraction models and provides valuable insights into their strengths and weaknesses. It serves as a foundation for further research in developing more efficient and effective network configurations and abstractions.
In computer networking, people used to use command-line interfaces to set up the network. But now there are other ways called network abstractions that make it easier. However, it's hard to compare which way is the best because people don't agree on what's most important. This makes it take longer for new ways to be used by everyone. One way to compare these different methods is by looking at how simple, efficient, and effective they are using a graph structure. Some existing methods like Kubernetes or Cisco Application Centric Infrastructure are good at storing network rules. But some methods in public clouds are too complicated. We need a fair way to compare all these methods so we can find the best one. Moreno's research does this and shows why it's important to compare them objectively. This research can help us make better networks in the future."
Definitions
- Traditional configuration methods: The old ways of setting up a network using command-line interfaces.
- Network abstractions: Different ways of setting up a network that make it easier than using command-line interfaces.
- Effectiveness: How well something works or achieves its goals.
- Lag in adoption: When something takes longer for everyone to start using because people don't agree on if it's good or not.
- Graph structure metrics: A way of measuring and comparing different things based on their connections and relationships.
- Network policies: Rules and settings for how a network should work.
- Excessive complexity: When something is too complicated or hard to understand.
Understanding Network Abstractions and Evaluating Their Effectiveness
In the field of computer networking, traditional configuration methods have relied on humans using command-line interfaces. However, in the past decade, various network abstractions have emerged as alternatives to manual configuration. Despite their availability, there is currently no objective way to compare these abstractions and determine their effectiveness. As a result, there is no consensus within the industry regarding the direction modern network abstractions should take, leading to a lag in their adoption.
A Comprehensive Evaluation Framework for Network Abstractions
To address this gap in understanding and facilitate informed decision-making regarding modern network abstractions, researcher Moreno has proposed a comprehensive evaluation framework based on graph structure metrics. This framework utilizes metrics derived from graph structures to evaluate the simplicity, efficiency, and effectiveness of different network abstraction models. By applying this framework to existing models such as Kubernetes or Cisco Application Centric Infrastructure (ACI) models versus public cloud-based abstractions like Amazon Web Services (AWS), Moreno finds that while some existing network abstractions are efficient for storing network policies others remain infrastructure-centric and suffer from excessive complexity.
The Need for Standardized Evaluation Methods
This research emphasizes the need for standardized evaluation methods to objectively assess different network abstraction models. By providing a comprehensive comparison framework based on graph structure metrics, Moreno's research contributes towards addressing this gap in understanding and facilitating informed decision-making regarding modern network abstractions. Overall, this study highlights the importance of objectively evaluating network abstraction models and provides valuable insights into their strengths and weaknesses. It serves as a foundation for further research in developing more efficient and effective network configurations and abstractions.
Computer networks have been traditionally configured by humans using command-line interfaces. Some network abstractions have emerged in the last 10 years, but there is no easy way of comparing them to each other objectively. Therefore, there is no consensus in the industry of what direction modern network abstractions should take, and the adoption of these abstractions lags as a consequence. In this paper I propose a comparison framework using metrics derived from graph structures to evaluate the simplicity, efficiency, and effectiveness of different network abstraction models. The result of this comparison is that while some of the existing network abstractions are quite efficient to store network policy (such as the Kubernetes or the Cisco Application Centric Infrastructure models), others (notably public cloud) are still very infrastructure-centric and suffer from excessive complexity.
