Browse Prior Art Database

Log Buffer Storage Optimisation

IP.com Disclosure Number: IPCOM000244654D
Publication Date: 2016-Jan-05
Document File: 4 page(s) / 48K

Publishing Venue

The IP.com Prior Art Database

Abstract

Optimised utilisation of memory allocated towards an event logging infrastructure, within a compute system.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 4

Log Buffer Storage Optimisation

Disclosed is a method to optimise the storage buffer utilisation in a multi user environment, where one to many shared memory buffer pool(s) are available for the purpose of event logging.

In a multi user environment, where multiple threads/applications are running on multiple

processors that interacts with hardware, an event logging component tends to be utilised that

provides services of event collection and logging, to all other threads/applications within the system. This helps in code failure debug and root cause identification. This task of event logging is achieved by using shared memory buffer pool(s) that are generally shared amongst the threads/applications in the compute system attempting to perform logging. To aid in flexible en-queue and de-queue of event log entries to/from the buffer pool(s), the buffer pool(s) are typically sliced into equal sized blocks. Each block within the memory buffer pool would house one event log entry.

In a memory constrained system, the aim is to utilise every byte of storage space allocated within the shared memory buffer pool(s), such that there is no lost event due to lack of storage memory in the shared memory buffer pool(s).

The Figure 1 below depicts an approach of facilitating event logging, where multiple shared memory buffer pool(s) are made available in the compute system, that

1


Page 02 of 4

Figure 1: Multiple buffer pool(s) with each buffer pool sliced up into equal sized slots

are shared amongst the various threads/applications within the multi user environment. Each shared memory buffer pool is sliced up into equal sized slots and each slot houses a single event log entry. The above design although having its advantages in ease of en-queue and de-queue of event logs, it also suffers 3 major drawbacks as listed below.


1. Not all of the event log entries that are housed in the 8K shared memory buffer pool or the 64K shared memory buffer pool would end up using the entire 8K or 64K bytes allotted for that

particular event log. This would lead to wastage of storage space within each slot, which will go unused.

2. The above setup as...