Browse Prior Art Database

Correction of ESA/390 Channel Measurement for Devices which Split Primary and Secondary Status

IP.com Disclosure Number: IPCOM000107745D
Original Publication Date: 1992-Mar-01
Included in the Prior Art Database: 2005-Mar-22
Document File: 2 page(s) / 80K

Publishing Venue

IBM

Related People

Meritt, AS: AUTHOR

Abstract

A method is disclosed to adjust the ESA/390* channel measurement data for time which is currently not captured. This is the time between channel end for the last command of a channel program and the device end signalling completion of the operation at the device. For devices which do this, such as non-synchronous DASD devices or buffered tape devices, this could amount to a large percentage of the actual service time and could therefore result in significant error in the device utilization and service time reported by measurement or monitor programs based strictly on the measurement data reported by the hardware.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Correction of ESA/390 Channel Measurement for Devices which Split Primary and Secondary Status

       A method is disclosed to adjust the ESA/390* channel
measurement data for time which is currently not captured. This is
the time between channel end for the last command of a channel
program and the device end signalling completion of the operation at
the device.  For devices which do this, such as non-synchronous DASD
devices or buffered tape devices, this could amount to a large
percentage of the actual service time and could therefore result in
significant error in the device utilization and service time reported
by measurement or monitor programs based strictly on the measurement
data reported by the hardware.

      The ESA/390 channel measurement facility provides 4
accumulators for every device being monitored:
      1.  queue time within the channel,
      2.  connect time, i.e., the time the device is actively
communicating with the channel,
      3.  disconnect time, i.e., the time during which an operation
is active at the device but the device is disconnected from the
channel such as when performing a seek, RPS positioning, or destaging
data from a buffer, and
      4.  queue time within the control unit.

      The device is considered in-use by this system for the sum of
(2)+(3); various calculations such as average service time or device
utilization can be derived from that sum. Thus, any loss of time
results in inaccuracies in these calculations.  For devices such as
non-synchronous DASD devices or buffered tape devices, the final
operation for write commands results in loss of the disconnect time,
which in fact is where most of the time for the operation really
occurs; loss of this portion of the time results in...