MSUs values differences between DCI Budget View and IBM SCRT

Products

• Dynamic Capacity Intelligence

Issue/Introduction

The Dynamic Capacity Intelligence (DCI) Budget View WebUI page shows monthly MSUs values that are higher than the MSUs values reported by IBM SCRT. For example:

	May-23	Jun-23	Jul-23	Aug-23	Sep-23	Oct-23	Nov-23	Dec-23	Jan-24	Feb-24	Mar-24	Apr-24
DCI BVW	681,491	633,846	666,835	708,484	634,459	669,767	630,290	640,581	668,414	621,101	655,202	648,940
IBM SCRT
SYS1	479,721	445,320	472,027	485,303	445,466	467,267	450,819	461,349	462,284	425,932	468,056	466,625
SYS2	203,141	188,998	196,111	224,478	190,397	203,744	180,732	180,494	206,193	196,408	188,541	183,521
Total	682,862	634,318	668,138	709,781	635,863	671,011	631,551	641,843	668,477	622,340	656,597	650,146
Delta	1,371	472	1,303	1,297	1,404	1,244	1,261	1,262	63	1,239	1,395	1,206

Environment

DCI 2.0 all releases

Resolution

The monthly difference between the two reports is quite small in percent, between 0.22% and 0.1%:

	May-23	Jun-23	Jul-23	Aug-23	Sep-23	Oct-23	Nov-23	Dec-23	Jan-24	Feb-24	Mar-24	Apr-24
DCI BVW	681,491	633,846	666,835	708,484	634,459	669,767	630,290	640,581	668,414	621,101	655,202	648,940
SCRT
SYS1	479,721	445,320	472,027	485,303	445,466	467,267	450,819	461,349	462,284	425,932	468,056	466,625
SYS2	203,141	188,998	196,111	224,478	190,397	203,744	180,732	180,494	206,193	196,408	188,541	183,521
Total	682,862	634,318	668,138	709,781	635,863	671,011	631,551	641,843	668,477	622,340	656,597	650,146
Delta	1,371	472	1,303	1,297	1,404	1,244	1,261	1,262	63	1,239	1,395	1,206
Delta Percent	0.2007726305	0.0744106268	0.1950195918	0.1827324203	0.2208022797	0.1853918937	0.1996671686	0.1966212921	0.0094244080	0.1990873156	0.2124590883	0.1854967961

 

This difference is expected, as you can read on the Budget View page of the DCI 2.0 documentation:
The actual values may vary from IBM SCRT values due to the rounding functions used in the SCRT. 

A lot of the variation between DCI and SCRT is due to the snapshot timing. Both the products use a 5 minute snapshot but they are not synchronised. It could be that there is 3+ minutes difference between the two snapshots for the same 5 minute interval. Usually things average out but if, for example, every LPAR is heading upwards quickly, whichever tool is taking the later snapshot will show higher values, which will be compounded by averaging. The variation is directly connected to the volatility and trend of the workload every 5 minutes.