The performance of the
Aggregation transformation is significantly improved through the use
of hash objects. In addition, input data to aggregations are now sorted.
This step improves the input and output processes, which are used
by the underlying SAS and operating system technologies that aggregate
IT data.
Moving averages and
moving standard deviations are now available as statistics that can
be included in aggregations. These moving statistics make SAS IT
Resource Management more useful because they enable IT organizations
to truly identify and establish baseline and threshold measurements
for the many performance metrics that they want to measure. These
statistics can also be used to monitor characteristics of the SAS
IT Resource Management system. For example, they can help monitor
the growth in the number of systems for which data is analyzed. They
can also monitor the volume of reports created by each SAS IT Resource
Management report job (if measures on those items are retained and
managed using SAS IT Resource Management).
SAS IT Resource Management
3.3 delivers support for calculating and including percentiles for
measurements within the IT data mart. Percentiles can be specified
and calculated over designated time periods for those aggregation
tables that are produced by the system. Percentile measurements in
SAS IT Resource Management enable IT organizations to quantify and
analyze utilization, availability, performance, and capacity characteristics
of IT infrastructure components. These measurements can be compared
with other components in the infrastructure so that IT organizations
can prioritize and resolve current and potential problems.
Cloning is
a new function that is available from the list of options that appears
when you right-click on an existing aggregated table of a job that
is open in the
Diagram tab of the
Job
Editor window. Many IT organizations establish separate
IT data marts for different IT resources based on their use by different
business organizations or to support production, testing, or development.
The cloning feature is useful because it allows Aggregation transformation
definitions to be created once and then copied and reused across other
IT data marts.
The
Summarized
Aggregation Table wizard is modified to simplify the
specification of an aggregation.
-
The
Select Analysis
Column page is eliminated. In all the pages where an
analysis column is requested, the list of all available columns is
displayed. For example, to specify percent change columns, you can
select the analysis columns from the list of class, ID, statistics,
and percentile columns.
-
The pages that specify additional
columns, such as statistics, percentiles, and so on, have a new, standardized
appearance. The information about the columns is displayed in a grid
format. The grid contains a row that describes the characteristics
of each column (class, ID, statistic, percentile, and so on) that
is specified. The row typically contains the name of the column, the
target column name, the target column description, and the target
column format. Additional information is displayed depending on the
type of column being described. For example, a row that describes
a statistic column displays its Weight By value, if weighting is specified.
In addition, a row that describes a percentile column displays its
Round To value.
Adding and deleting
columns from the aggregation table is accomplished by clicking the
New and
Delete buttons,
respectively.
-
The MACHINE column is removed from
the class list of these Key Metrics Disk aggregations:
Note: If the machines that are
being monitored use the preceding adapters and are attached to a storage
area network, then each disk within that storage area network appears
as if it were installed locally on that machine. Aggregating such
data from the perspective of the host system is of little, if any,
value. Therefore, the MACHINE column has been removed from the class
list of the Key Metric Disk aggregations for these data sources.
-
The retention period for the DayHour,
DayShift, and DayShiftHour aggregations has changed from 92 to 45
days. The retention period remains the same for other aggregation
tables.