Packet Metadata Compaction for Small-Message Coalescing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The metadata overhead in network packets, especially for small messages, significantly reduces network efficiency by consuming bandwidth that could otherwise be used for payload data, leading to low payload bandwidth efficiency.
Innovation Solution
Implementing per-message metadata compaction by representing metadata values with smaller codes and configuring senders and receivers to use these codes, allowing multiple sequential messages to share a single network header, thereby reducing redundant metadata transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple small messages are sent separately, then each message can be transmitted independently, but metadata overhead is duplicated for each message reducing network efficiency
Solution Approach 1:
The patent merges multiple individual message transmissions into a single coalesced packet. Multiple small messages that would normally be sent separately are combined into one packet with a shared network header, eliminating redundant metadata transmission and improving bandwidth efficiency while maintaining the ability to process individual messages at the destination.
Solution Approach 2:
The patent segments the metadata into two parts: a shared network header containing common metadata fields (source/destination addresses, protocol information) that is transmitted once for all coalesced messages, and individual message metadata (payload data) that is transmitted for each message. This segmentation reduces redundant metadata while preserving message independence.
2Loss of energy
If messages are coalesced into a single packet, then metadata overhead is reduced by sharing network headers, but individual messages still carry redundant metadata fields
Solution Approach 1:
The patent applies different metadata handling strategies to different parts of the packet. The network header metadata is shared across all coalesced messages, while individual message metadata is selectively transmitted only when necessary. This local differentiation optimizes the balance between reducing overhead and maintaining message identification capabilities.
3Loss of information
If per-message metadata is transmitted in coalesced packets, then message identification is possible, but bandwidth is consumed that could be used for payload data
Solution Approach 1:
The patent extracts the common metadata fields from individual message transmissions and places them in a shared network header. Only the essential message identification metadata that cannot be shared is transmitted for each individual message, minimizing bandwidth consumption while preserving the ability to identify and process each message correctly.
Data Source
AI summary
Examples described herein relate to coalescing one or more messages into a coalesced message and representing one or more fields of the metadata of the one or more messages using one or more codes, wherein at least one of the one or more codes uses fewer bits than that of original metadata fields to compact the metadata fields. In some examples, the metadata includes at least one or more of: a target processing element (PE) number or identifier, message length, operation to perform, target address where to read or write data, source PE number or identifier, initiator address in which to write result data, or message identifier.