MSRP Layer-2 Bandwidth Reservation for Bridged Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current protocols for resource allocation in bridged computer networks, such as RSVP, are ill-suited for Layer-2 networks as they operate at Layer-3 and do not account for the specific requirements of Layer-2 devices like bridges and switches, leading to inefficiencies in bandwidth reservation and congestion management.
Innovation Solution
The development of the Multiple Stream Reservation Protocol (MSRP) that allows bridges to allocate resources for data streams by defining new protocols like Multiple Payload Registration Protocol (MPRP) and MSRP, enabling efficient bandwidth reservation and resource management within Layer-2 networks, even in the presence of shared media and multiple streams.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If RSVP protocol is used for resource allocation, then bandwidth reservation is achieved, but it is incompatible with Layer-2 bridged networks
Solution Approach 1:
The patent introduces a Layer-2 adaptation mechanism that acts as an intermediary between RSVP (Layer-3) and bridged networks (Layer-2). The adapter translates RSVP reservation requests into Layer-2 compatible format, allowing bandwidth reservation functionality to work in bridged networks without requiring end systems to be aware of Layer-2 specifics.
Solution Approach 2:
The patent moves the resource allocation function from Layer-3 (network layer) to Layer-2 (data link layer) by implementing reservation functionality within the bridge itself. This dimensional shift allows the protocol to operate natively in Layer-2 bridged networks while maintaining compatibility with higher-layer applications.
2Productivity
If multiple streams share network resources, then network utilization increases, but congestion and packet loss increase
Solution Approach 1:
The patent segments network resources by creating virtual channels within the bridge, where each data stream is assigned to a specific virtual channel with guaranteed bandwidth allocation. This segmentation allows multiple streams to coexist without interfering with each other, maintaining both high network utilization and reliable packet delivery.
Solution Approach 2:
The patent implements preliminary resource allocation by reserving bandwidth for each stream before data transmission begins. The bridge pre-allocates buffer space and transmission slots for each virtual channel, ensuring that congestion and packet loss are prevented before they occur during actual data flow.
3Reliability
If bandwidth is reserved for specific streams, then QoS is improved, but network flexibility decreases
Solution Approach 1:
The patent implements dynamic resource allocation where the bridge can adjust bandwidth allocation for different virtual channels based on current network conditions and stream requirements. The system allows streams to be added, removed, or modified dynamically while maintaining QoS guarantees for active streams, thus preserving both reliability and flexibility.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In one embodiment, a talker device may issue talker registrations to bridges of a network domain for a stream, the talker registration having at least a bandwidth requirement and a state of the talker registration as either offering or failed. Also, a listener device may issue listener registrations for a stream, the listener registration having at least a state of the listener registration as one of asking, asking-failed, ready, or ready-failed. In response to receiving a talker registration and listener registration for the same stream, a bridge of the network domain may then attempt to allocate resources for the stream if the bridge is on a path of the stream between the talker device and the listener device. The bridge may then notify, via respective states of the talker and listener registrations, the talker device and the listener device of whether resources have been allocated for the stream.