Browse Prior Art Database

Single Non-Volatile Memory Services Multiple NVM Requirements with Redundancy to Simplify Program Updates and Limit Field Corruption

IP.com Disclosure Number: IPCOM000182465D
Original Publication Date: 2009-Apr-30
Included in the Prior Art Database: 2009-Apr-30
Document File: 4 page(s) / 31K

Publishing Venue

IBM

Abstract

Today's servers have multiple, separate NVMs (non-volatile memories). These include the flash BIOS (Basic Input/Output System) and various local EEPROMs (Electrically Erasable Programmable Read-Only Memories) connected to CPLDs (Complex Programmable Logic Devices), FPGAs (Field Programmable Gate Arrays), controllers, hubs, etc. Each requires a separate method and code to program, and only the BIOS is typically backed up with either a locked boot block or duplicate image. Field data shows that there are a significant number of motherboards that are returned for service because one of their NVM's is corrupted. Disclosed is a method of combining all these devices into a single NVM that can serve as primary and backup memory for all the server's needs and be efficiently managed.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 62% of the total text.

Page 1 of 4

Single Non-Volatile Memory Services Multiple NVM Requirements with Redundancy to Simplify Program Updates and Limit Field Corruption

This idea eliminates field corruption and enables a single method of updating a server's NVM images by consolidating the NVMs to a single double-size memory module behind an arbitration point that appears to each NVM-using device as a single, separate, dedicated component. The idea consists of these ingredients:

1) Non-Volatile Memory


A NVM with at least twice the capacity of the sum of the capacities of the separate NVMs required. The access bus can be a higher speed to reduce the latency to the using devices.

2) Bus Virtualization Circuit


This circuit consists of bus converters translating to a Common Bus (CB) and steered sequentially to the single NVM by the Bus Arbitor. Registers within the Bus Arbitor would determine the address of the common NVM as seen by each using device. This register set could be accessed thru a preferred device bus, or thru a separate bus or pins. Another bus converter translates CB back to the preferred bus of the NVM.

3) Corruption Detection Circuit


This can be one of any number of conventional schemes used to detect that the image is not valid. A watchdog timer would be one example. If the timer is not reset by software periodically, it times out and switches over to the backup image.

Additional conventional in-band and out-of-band circuits can be added to this...