ATCA Blade Chassis Management via IPMB Multi-Master Protocol
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current ATCA chassis management systems require dedicated network blades and a complex shelving manager card, leading to inefficient slot occupancy, high costs, and complex networking in large-scale installations, especially in configurations without network blades.
Innovation Solution
Implementing a point-to-point, peer-to-peer chassis management configuration using a non-standard multi-master protocol on the IPMB bus, where all productive blades manage themselves and share information, eliminating the need for a shelving manager card and optimizing slot usage without physical modifications to the backplane.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If dedicated network blades and shelving manager card are used for chassis management, then chassis management functionality is ensured, but slot occupancy efficiency decreases and system complexity increases
Solution Approach 1:
The invention extracts the dedicated chassis management components (network blades and shelving manager card) from the system and replaces them with a simplified approach where any productive blade can perform management functions. This eliminates the need for specialized management hardware while maintaining chassis management capabilities through the existing IPMB bus and multi-master protocol.
Solution Approach 2:
The invention makes all productive blades universal by enabling them to perform both productive work and chassis management functions. Any blade can act as a management node, eliminating the need for dedicated network blades and shelving manager cards. This multi-functionality increases slot occupancy efficiency while reducing system complexity.
2Reliability
If dedicated network blades are used for chassis management, then management functionality is provided, but the number of productive blades available decreases
Solution Approach 1:
The invention enables all blades to serve dual purposes: productive operations and chassis management. By allowing any productive blade to perform management functions through the multi-master protocol on the IPMB bus, the system eliminates dedicated network blades, thereby maximizing the number of productive blades available while ensuring management functionality is maintained.
Solution Approach 2:
The invention removes dedicated network blades from the system architecture, extracting their management functions and distributing them across all productive blades. This extraction increases productive blade availability while maintaining management capabilities through the decentralized multi-master approach.
3Reliability
If shelving manager card is used for chassis management, then centralized management control is achieved, but cost and device complexity increase
Solution Approach 1:
The invention extracts the centralized shelving manager card from the system and replaces it with a decentralized management approach. Management control is distributed across multiple productive blades that communicate through the IPMB bus using multi-master protocol, eliminating the need for a dedicated management card while maintaining control functionality.
Solution Approach 2:
The invention enables blades to self-manage chassis operations through the multi-master protocol. Each blade can initiate and respond to management commands autonomously, eliminating the need for a centralized shelving manager card. This self-service approach reduces device complexity and cost while maintaining management control.
4Reliability
If network blades are used for chassis interconnection, then internal networking is ensured, but slot occupancy and cost efficiency decrease
Solution Approach 1:
The invention makes all blades universal nodes that can perform both productive operations and internal networking functions. By enabling any productive blade to participate in chassis interconnection through the multi-master protocol on the IPMB bus, the system eliminates dedicated network blades, thereby maximizing slot occupancy efficiency while ensuring internal networking capability.
Data Source
Figure 1~2b
Figure 3~4
AI summary
The chassis comprises interconnected blades (10) in a network (36), a control board (16) for the power supply and ventilation units, and a serial bus (30) dedicated to receiving and disseminating information relating to status data or electrical or physical parameters measured or detected on the blades or the chassis. Chassis management is distributed across the blades in a non-hierarchical point-to-point configuration, using a protocol compliant with the CAN specification. Each blade analyzes blade and chassis information and generates management commands applied to the bus for dissemination to the other blades and the master control board. The hardware characteristics of the blades and the backplane board remain compliant with the ATCA specification, while the bus hardware characteristics are those of an IPMB-type bus.