Browse Prior Art Database

Algorithm for Calculating Optimum Receive Buffer Size and Number in Token Ring Adapter

IP.com Disclosure Number: IPCOM000107154D
Original Publication Date: 1992-Jan-01
Included in the Prior Art Database: 2005-Mar-21
Document File: 2 page(s) / 57K

Publishing Venue

IBM

Related People

Scherer, PC: AUTHOR [+2]

Abstract

This article describes an algorithm which may be used to calculate optimum Receive Buffer Size and Number of Receive Buffers for all supported IBM Token Ring Adapters.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 60% of the total text.

Algorithm for Calculating Optimum Receive Buffer Size and Number in Token Ring Adapter

       This article describes an algorithm which may be used to
calculate optimum Receive Buffer Size and Number of Receive Buffers
for all supported IBM Token Ring Adapters.

      The algorithm starts with a Receive Buffer Size of 512 bytes
and calculates the amount of memory required for this buffer size.
The innovative feature of this algorithm is its ability to adjust for
an additional amount of memory required for configurations with a
large number of links.

      The amount of memory required by the Receive Buffers for a
configuration is calculated using the following equation:
      A = R * int((X - 6)/(R - 8) * N * ((P * L)
+ 1) + O) where
      A is the Memory Requirement in bytes
      R is the Receive Buffer Size in bytes
      X is the Transmit Buffer Size in bytes
      N is the Number of Transmit Buffers P is the Percentage of
Links
      Simultaneously Active (default = 0.025) L are the Maximum
Number of Link
      Stations O is the Paging Overhead (IBM Token Ring 16/4
Adapters=3; Others=0) int means to round function to the closest
integer

      If the memory requirements for the proposed maximum Receive
Buffer Size of 512 bytes exceed the available memory on the Token
Ring Adapter, the next smaller Buffer Size of 504 bytes is selected.
As long as there is no memory fit, the proposed Receive Buffer Size
is decremented...