IBM 5695-DF1, 5655-DB2 manual Cache effectiveness Overall Summary, Space Utilization Summary

12.2.2.2 Cache effectiveness Overall Summary

Figure 73 on page 153 shows an extract of these reports for both RVAs that contains information similar to RMF, but with more details on caching algorithms, and also explains the origin of the observed disconnect times. The I/O demand the host submits to the RVAs and the staging effectiveness are analyzed:

•Analysis of host I/Os. Both RVAs receive a similar I/O activity demand which is read-only. That explains the high value of the READ RATIO field, which is the read-to-write ratio. Moreover, the sum (read-plus-write) is larger than number of I/Os. There are probably more than one read per I/O.

•Hit ratios are more than 99%. That is an indication that practically all reads find their addressed tracks into the cache. This situation is current when a sequential pre-staging algorithm anticipates the I/O read demand.

•Staging activity (measured in number of tracks per second) is 2.4 times the I/O activity:

(113.2 + 103 ) / ( 45.7 + 44.5 ) = 2.4

There are 2.4 tracks staged per I/O: This intensive staging activity explains the observed disconnect time.

XSA/REPORTER				CACHE EFFECTIVENESS OVERALL SUMMARY							18FEB1999			17:32:04
SUBSYSTEM NAME: 20395	(CACHE SIZE:		1024 MB		NVS SIZE:		8 MB)
SUBSYSTEM	READ	WRITE		I/O	READ		READ	WRITE	I/O	DFW	STAGE	HITS/		LOW	TRACK
SUMMARY	PER SEC	PER SEC		PER SEC	RATIO	HIT %		HIT %	HIT %	CONSTR	PER SEC	STGE	REF CT		OCCUP
	-------	-------		-------	-----	-----		-----	-----	------	-------	-----	------		------
PROD PARTITION	53.8	0.0		45.7	61329		99.3	100.0	99.3	0.0	113.2	0.5		73.7
OVERALL TOTALS	53.8	0.0		45.7	61329		99.3	100.0	99.3	0.0	113.2	0.5		73.7	25050
SUBSYSTEM NAME: 22897	(CACHE SIZE:		1280 MB		NVS SIZE:		8 MB)
SUBSYSTEM	READ	WRITE		I/O	READ	READ		WRITE	I/O	DFW	STAGE	HITS/		LOW	TRACK
SUMMARY	PER SEC	PER SEC		PER SEC	RATIO	HIT %		HIT %	HIT %	CONSTR	PER SEC	STGE	REF CT		OCCUP
	-------	-------		-------	-----	-----		-----	-----	------	-------	-----	------		------
PROD PARTITION	50.1	0.0		44.5	57064		99.9	100.0	99.9	0.0	103.0	0.5		73.5
OVERALL TOTALS	50.1	0.0		44.5	57064		99.9	100.0	99.9	0.0	103.0	0.5		73.5	22086

Figure 73. Case Study IXFP Cache Effectiveness Overall Extract

12.2.2.3 Space Utilization Summary

Figure 74 on page 154 shows the extracts of space utilization reports for both RVAs. Table 27 on page 153 summarizes different space utilization by both RVAs. It also shows how, in this balanced configuration with an evenly distributed workload and with consistent service times, the NCLs are not homogenous.

Table 27. RVA Space Utilization Comparison

Subsystem	20395	22897
Net Capacity Load	56.4%	25.2%
% Functional Capacity Stored	28.1	5.4
Compression Ratio	3.1	1.9

The partitions of the table spaces into RVA 22897 contain data already compressed by CPU. From performance and transfer bandwidth points of view (KB per second), both RVAs have similar appeareances.

Case Study 153