Seamless High Bandwidth Lane Addition Without Buffering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In high-bandwidth communication networks, seamlessly adding new communication lanes without disrupting fixed delay transmissions is challenging, particularly during the training phase, as existing methods require buffers to maintain delay consistency.

Innovation Solution

The method involves exchanging indications between devices to add new high-bandwidth lanes using 8b/10b code words with fixed delays, sending idle and synchronization sequences, and utilizing known non-idle sequences for deskewing, allowing seamless integration without buffers and maintaining continuous data transmission.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If buffers are used to maintain fixed delay during training phase, then delay consistency is preserved, but device complexity and power consumption increase

Engineering Contradiction:
Improvedelay consistencyVSAvoidbuffer requirement
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent changes the operational parameters of communication lanes during the training phase. Specifically, it transitions lanes from a data transmission mode with fixed delay to a training mode where lanes can operate independently without buffer constraints. This parameter change allows the system to achieve lane training without requiring buffers, thereby reducing device complexity while maintaining overall system reliability through coordinated transition protocols.

Inventive Principle:
Principle #35Parameter changes

2Use of energy by moving object

If all inter chip communication lanes are turned off to save power, then energy consumption is reduced, but network adaptability decreases when lanes need to be dynamically added

Engineering Contradiction:
Improvepower consumptionVSAvoidlane addition capability
Core Design Contradiction:
Use of energy by moving objectVSAdaptability or versatility

Solution Approach 1:

The patent implements preliminary action by establishing a standardized training sequence and synchronization protocol before lanes are activated. When lanes need to be dynamically added, the system can quickly initialize them using pre-defined training sequences without requiring complex buffer management or interrupting existing traffic. This preliminary preparation enables efficient lane addition while maintaining power-saving mode for inactive lanes, thus preserving both energy efficiency and network adaptability.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If new high bandwidth lanes are added during operation, then network capacity is improved, but transmission stability is disrupted during the training phase

Engineering Contradiction:
Improvebandwidth capacityVSAvoidtransmission stability
Core Design Contradiction:
ProductivityVSStability of the object's composition

Solution Approach 1:

The patent applies segmentation by separating the lane training process from the data transmission process. Existing active lanes continue to operate with fixed delay and stable transmission, while new lanes undergo training in parallel using dedicated training sequences. This segmentation allows bandwidth capacity to be improved through lane addition without disrupting the stability of ongoing transmissions, as each lane group operates independently during the transition.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS9594719B2Seamless addition of high bandwidth lanes
Publication Date: 2017.03.14 VALENS SEMICON
  • US9594719B2 patent drawing
  • US9594719B2 patent drawing
  • US9594719B2 patent drawing

AI summary

Seamless addition of high bandwidth lanes, including the steps of: sending, by a master, an idle sequence using 7b/10b code words over new high bandwidth lanes in parallel to sending and receiving 8b/10b data with a fixed delay over master-to-slave (m2s) and slave-to-master (s2m) active high bandwidth lanes; sending in parallel a synchronization sequence and a known non-idle sequence during an inter packet gap; utilizing, by the slave, the known non-idle sequence for deskewing the new high bandwidth lanes; and sending, by the master, a transition sequence over both the m2s active high bandwidth lane and the new high bandwidth lanes, and immediately thereafter the master is ready to transmit high bandwidth data using 8b/10b code words over both the m2s active high bandwidth lane and the new high bandwidth lanes.