NDN Flow Balance with Variable Data Object Sizes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Named Data Networking (NDN) protocols constrain data messages to be small, leading to inconvenient data object sizes for many applications and increased computational overhead due to fragmentation, while failing to handle a dynamic range of data sizes effectively, resulting in resource allocation challenges and high retransmission costs for large data objects.
Innovation Solution
Incorporating an 'Expected Data Size' control field in Interest packets to accurately allocate resources based on data object size, allowing for a larger dynamic range of data sizes and maintaining flow balance by ensuring fair resource allocation between small and large data objects, and using a 'Data Too Big' message to handle mismatches between expected and actual data sizes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If flow control/balance is maintained using simple bookkeeping to track outstanding Interest messages, then flow balance is achieved, but data object sizes are constrained to be small and near constant
Solution Approach 1:
The patent introduces a new parameter 'Expected Data Size' in the Interest message to enable dynamic adaptation to different data object sizes. This parameter allows the system to maintain flow balance while accommodating variable data sizes, resolving the contradiction between simple bookkeeping and size adaptability.
Solution Approach 2:
The system transitions from static flow balance (fixed MTU-sized data objects) to dynamic flow balance that adapts to actual data sizes. The Interest message dynamically specifies expected data size, allowing the network to adjust resource allocation accordingly while maintaining fairness.
2Device complexity
If data messages are constrained to be small (near MTU size), then flow balance can be maintained with simple bookkeeping, but data flow is constrained and computational overhead increases due to fragmentation
Solution Approach 1:
By introducing the 'Expected Data Size' parameter in Interest messages, the system enables efficient data transmission without fragmentation. Network nodes can allocate resources based on actual data sizes, improving productivity while keeping the flow control mechanism relatively simple.
Solution Approach 2:
The Expected Data Size is specified in advance in the Interest message before data transmission begins. This preliminary information allows network nodes to pre-allocate appropriate resources, avoiding fragmentation and improving data flow efficiency.
3Device complexity
If simple bookkeeping is used to track outstanding Interests, then implementation is simple, but large data objects experience high retransmission costs and resource allocation challenges
Solution Approach 1:
The patent extends the simple bookkeeping approach by adding the 'Expected Data Size' parameter to Interest messages. This enhancement allows the system to handle large data objects effectively while maintaining implementation simplicity, as nodes still use basic bookkeeping augmented with size information for better resource allocation.
Data Source
AI summary
A network device in a Named Data Networking (NDN) network receives an Interest from a consumer including a name and an Expected Data Size of data requested through the Interest. The network device forwards the Interest along a path to a producer of the data based on the name. As a result, the network device receives data that has traversed the path in reverse and satisfies the forwarded Interest. The network device determines an actual data size of the received data. The network device compares the actual data size to the Expected Data Size. If the actual data size is greater than the expected data size, and if a level of traffic congestion associated with forwarding the received data to the consumer is below a threshold, forwarding the received data to the consumer along a path based on the name of the data.