2. USAGE CONSIDERATIONS › 2.3 Use Within the Information Structure › 2.3.3 Service Level Analysis › 2.3.3.2 Service Levels: Throughput

2.3.3.2 Service Levels: Throughput

 
Throughput constraints and capabilities are of primary
concern to the network analyst.  As in mainframe capacity
planning, the network analyst must consider the various
workloads that affect the data center.  Workload can
conveniently be divided into two broad classes:  logical
workloads and physical workloads.
 
 
LOGICAL WORKLOADS
 
Logical workloads are directly related to the end user.
Planning for new systems or major configuration changes is
usually concerned initially with logical workloads.  Logical
workloads can be subdivided into two categories:  application
workloads and organization (business) unit workloads.
 
Application workloads are identified in the following files:
 
o  SNTPSY - NPM System Activity File
o  NVSLSY - NLDM System Activity File
o  SNTNSP - CA NetSpy Application File
 
Organizational unit workloads can be identified via the use
of account code data elements in the following files:

o  NCP Network Accounting (SNTNAC) File
o  CA NetSpy Terminal (SNTNSS) File
o  NPM User Activity (SNTPSU) File
o  NLDM Session Accounting (NVSNSA) File
o  NLDM Session Connectivity (NVSNSC) File
o  NLDM Response Time (NVSRTM) File
o  NLDM Route (NVSRTE) File

 
Sample account code definition members are included with this
product, and the use of account codes is discussed in Section
7.2.  However, it is your responsibility to verify that the
account code structure meets the requirements of your
data center, and, if not, to make any necessary
modifications.  In this way, you are assured the greatest
flexibility in defining an account code structure that will
satisfy your network analysis and reporting requirements.
 
 
PHYSICAL WORKLOADS
 
Physical workloads are directly related to network topology
and may only incidentally be related to the end user.
Localized network performance considerations such as load
balancing and NCP parameter selection are most often
concerned with physical workloads.  Physical workloads can be
divided into two categories:  boundary workloads and backbone
workloads.
 
The sequence/summary variable SLULINK in the NPM Link
Activity (SNTNPL) File and the NLDM Link Activity (NVSNLL)
File contains the name of the peripheral link to which the
secondary logical unit's physical unit is attached.  Network
boundary (peripheral link) workloads may thus be identified
directly.  Coincidentally, they may also be identified in the
CA NetSpy Terminal (SNTNSS) File, and in the NLDM Service
(SNTNSV) and the NLDM System Activity (NVSLSY) Files,
depending on the definition of performance classes for
NetView/NLDM and RTM classes for CA NetMaster NTS.
 
Performance problems related to boundary workloads are
limited in scope because they are localized to a specific
peripheral link.  However, if the link attaches a distributed
processor, a significant number of end users may be affected.
 
Backbone workloads may be identified by summarizing
utilization data to the level of the primary and secondary
logical unit's subareas (note that the CA MICS summarization
facility, explained in the CA MICS Standard Reporting Guide,
should always be utilized for any summarization of CA MICS
files).  This yields the traffic load between subarea "x" and
subarea "y".
 
If you are processing CA NetSpy Virtual Route Activity data,
you will find information about both SNA backbone virtual
route traffic volumes and response times in the CA NetSpy
Virtual Route Activity (SNTNVR) File.
 
Performance problems related to backbone workload constraints
can have far-reaching effects in the network and should be
closely tracked.  Relatively small changes in the traffic mix
(message lengths, chaining characteristics, etc.) can cause
significant increases in the utilization level of the
backbone communications control units.
 
The following sections suggest key data elements, exception
conditions, reports, and sample analysis that may be helpful
in the management of network throughput levels:
 
   1 - Throughput Level Indicators
   2 - Throughput Level Exceptions
   3 - Throughput Level Reports
   4 - Throughput Level Analysis