Bookshelf v7.5: Planning the Distribution of Your Database Objects

Regardless of the RDBMS you implement, and your chosen disk arrangement, be certain that you properly distribute the following types of database objects:

In most implementations, the Siebel tables listed in Table 29 and their corresponding indexes tend to be either the most commonly used, or they have the potential to be large in some, or in all, deployments. For example, the tables S_EVT_ACT, S_CONTACT, and S_ORG_EXT are large in all sizeable deployments of Siebel eBusiness Applications.

Table 29. Most Frequently Used and Largest Siebel Tables
Table Names
`S_ACCNT_CHRCTR`	`S_INVOICE`
`S_ACCNT_CO_MSTR`	`S_INVOICE_ITEM`
`S_ACCNT_POSTN`	`S_INV_LGR_ENTRY`
`S_ADDR_ORG`	`S_OPTY_POSTN`
`S_ADDR_PER`	`S_OPTY_PROD`
`S_ASSET`	`S_OPTY_TERR`
`S_CALL_LST_CON`	`S_OPTY_POSTN`
`S_CON_CHRCTR`	`S_ORG_EXT`
`S_CON_TERR`	`S_ORG_TERR`
`S_ACCNT_CHRCTR`	`S_PARTY`
`S_CRSE_TSTRUN`	`S_PARTY_PER`
`S_CRSE_TSTRUN_A`	`S_PARTY_REL`
`S_CS_RUN`	`S_PARTY_RPT_REL`
`S_CS_RUN_ANSWR`	`S_POSTN_CON`
`S_CTLGCAT_PATH`	`S_INVC_ITM_DTL`
`S_CYC_CNT_ASSET`	`S_INVLOC_ROLLUP`
`S_DNB_CON_MRC`	`S_POST_CON`
`S_DNB_ORG`	`S_PROC_REQ`
`S_DNB_ORG_SIC`	`S_PROD_BASELINE`
`S_DNB_UPDATE`	`S_PROD_CONSUME`
`S_DOCK_INIT_ITEM`	`S_PROD_SHIPMENT`
`S_DOCK_TXN_LOG`	`S_PROD_TARGET`
`S_DOCK_TXN_LOGT`	`S_QUOTE_ITEM`
`S_DOCK_TXN_SET`	`S_SRM_REPLY`
`S_DOCK_TXN_SETT`	`S_SRM_REQUEST`
`S_ESCL_ACTN_REQ`	`S_SRM_REQ_PARAM`
`S_ESCL_LOG`	S_SRV_REQ
`S_ESCL_REQ`
`S_EVT_ACT`
`S_EXP_ITEM`
`S_EXP_RPT`
`S_EXP_RPT_APPR`
`S_IC_CALC`
`S_IC_CALC_IT`
`S_IC_CMPNT_EARN`
`S_IC_TXN`
`S_IC_TXN_IT`
`S_IC_TXN_POSTN`

These tables and indexes should be separated across devices. As a general rule, indexes should be in a different tablespace and, if possible, on different physical devices from the tables on which they are created.

Planning Database Distribution if a RAID Disk Array Is Not Used

If a RAID device is not in use, even if space is at a premium, you must separate the indexes whose names end with _P1 from the tables on which they are created. These tables are heavily used in join operations.

If you will make frequent use of Siebel Enterprise Integration Manager (EIM), you may want to put the interface tables and indexes (names starting with EIM_) on different devices from the Siebel base tables. Both tables are accessed simultaneously during EIM operations.