Combined Transaction Processing and Reporting Benchmark - Benchmarking Transaction and Analytical Processing Systems - page 67

the past. Another aspect includes process optimization and customer satisfaction
questions, for example, end-to-end order processing time, order completion, that is,
the percentage of orders completed in time, and reducing the average amount of
accounts receivable by analyzing and speeding up cash collection. In the following,
the conceptual data model containing the most important entities that take part
in the order-to-cash process is introduced.

4.2.2 Conceptual Data Model and Database Schema

The entities taking part in the order-to-cash process can be classified according to
the process step they are primarily taking part in. The process steps are (a) ordering,
(b) delivery, (c) billing, and (d) accounting. Figure 4.4 gives an overview of the
entities within this scenario.
Master data, which is depicted in gray shading includes data about the products
offered, sales organizations as administrative units of the company that offer
products, for example in a specific area or line of business, business partner
data, location and contact information of business partners and sales organizations.
Business partner is a general term for parties connected with the sales process at
the customer's side. Sold-to-party and ship-to-party are typical examples. All other
data entities contain transaction data, which is changed frequently during business
processing. This includes the sales orders that reference specific products, their
deliveries, invoices for delivered orders and accounting information, which covers
the financial view of the sales process.
Figure 4.5 depicts an overview of the database schema as taken from a real
enterprises transactional system. It shows the excerpt of the tables used in the order-
to-cash process, their basic relationships, and the most important attributes. The
number of columns within the tables of this original schema is quite large, varying
between 5 and 327 columns. The total number of columns is 2,316, at least an
order of magnitude more than in the standard benchmarks discussed in Sect. 3.2 .
One reason for this large amount of attributes is the occurrence of redundant values
and pre-computed aggregates in the original data set. This is an optimization so far
used to speed up data access and to avoid joins. Because of the large number of
columns in the database schema, only those accessed most often in the transactions
and queries, which are going to be introduced in the following sections, are depicted
in Fig. 4.5 .
A general pattern, which can be derived from this database schema, is that each
set of tables contains a header and line item table. This conforms to the setup of
a database in second normal form. Header tables contain only general information,
e.g., creation date and information applicable to the entire set of line items belonging
to it like the buyer who initiated the sales order. General information of the header
depends on an identifier, e.g. order ID as the identifier of a sales order header.
Detailed information stored in the line item tables is fully dependent on the entire