Hardware Management Server Quick Inventory via Rack Hashing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The hardware management system (HMS) in software-defined data centers (SDDC) faces significant time constraints during hardware discovery and inventory processes, particularly in multi-rack environments, due to the need for extensive data querying and parsing, resulting in prolonged boot-up times and inventory completion.
Innovation Solution
Implementing a method where the HMS creates a hash database of rack attributes, allowing for rapid identification of matching inventory hashes to load pre-existing rack inventory data instead of performing a full discovery, and updates this data based on sensor error logs, significantly reducing the time required for inventory and boot-up processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If HMS performs full hardware discovery and inventory querying for each rack, then complete and accurate hardware inventory is achieved, but the time required for boot-up and inventory completion increases significantly
Solution Approach 1:
The system performs hardware discovery and inventory querying in advance, storing the collected data in a cache before the HMS actually needs it. This preliminary action allows the system to have inventory data ready when boot-up occurs, eliminating the need to perform time-consuming discovery processes during startup while maintaining inventory accuracy.
Solution Approach 2:
The system creates copies of hardware inventory data by hashing rack attributes and storing these hash values in a cache. Instead of querying all hardware details during boot-up, the system copies relevant inventory information from the cache based on hash matches, significantly reducing the time required while preserving data accuracy.
2Quantity of substance
If HMS performs comprehensive data querying and parsing for multi-rack inventory, then complete inventory coverage is achieved, but the complexity and time required for processing increases
Solution Approach 1:
The system extracts only the essential rack attributes needed for identification and creates hash values from these attributes. Instead of processing and storing all possible hardware data during discovery, the system takes out only the critical identifying information, reducing processing complexity while maintaining the ability to retrieve complete inventory data when needed.
Solution Approach 2:
The system transforms rack attributes into hash values, changing the parameter representation from detailed hardware specifications to compact hash codes. This parameter transformation enables efficient storage and retrieval of inventory data without requiring complex processing of the original detailed attributes during normal operations.
3Ease of operation
If HMS is stateless and performs rack discovery every time it boots up, then the system maintains simplicity and statelessness, but the boot-up time increases due to repeated full discovery processes
Solution Approach 1:
The system performs hardware discovery and creates inventory hash caches in advance, before the HMS needs to boot up. This preliminary action stores the results of what would otherwise be required during each boot-up sequence, allowing the HMS to remain stateless while dramatically reducing boot-up time by loading pre-computed inventory data from the cache.
Data Source
AI summary
A quick hardware inventory of a software-defined data center (SDDC) multi-rack is created. A hardware management server (HMS) receives an imaging of a first rack and performs a node hash of attributes of the imaging of the first rack. A hash database is then checked for an inventory hash that matches the node hash of the first rack. If a match is found, a rack inventory is loaded from a persistent rack data inventory correlated with the inventory hash instead of the HMS performing a discovery and inventory of the first rack. In addition, a sensor error log of the first rack is checked for any errors and the rack inventory is modified based on any errors in the sensor error log. The rack inventory is then saved into the server switch object cache of the HMS. The steps are repeated by the HMS for each additional rack.