Process Data Warehouse

Wednesday, November 9, 2011

Related Work - Mangisengi

Mangisengi et al. present an activity warehouse model [11]. Thereby, they motivate the warehouse by introducing requirements and measurements derived by BAM. Consequently, they introduce a simple warehouse model for storing audit trail logs and give time, organization, and process as dimensions.

[11].
Mangisengi, O. Pichler, M., Draheim, D.: Activity warehouse. In: ICEIS. pp. 137-144 (2007)

Tuesday, November 8, 2011

What and Why?

The majority of enterprises begin to capture their processes via business process
modeling. Thereby, they get an overview of the steps necessary to deliver their
offered services. Nowadays, lots of companies engage work
ow management systems
(WFMSs) to automate their business tasks. As a result, they reduce cost
and increase the e ciency of their processes. Although a WFMS gathers all
process-related information, it usually o ers limited analysis capabilities. It only
gives process instance counts and some aggregations like average cycle time.
Consequently, deeper process insights are hard to generate and, as consequence,
require manual actions from process analysts. Nevertheless, these insights
are getting more value in today's businesses. From the strategic view on
the e ciency of processes enterprise move to the real-time monitoring of business
activities.
Facing this upcoming development, this blog discusses the requirements
in detail to motivate the process data warehouse (PDW). The PDW merges
diverse ongoing approaches to collecting, processing, and storing of audit trail
logs. Thereby, this blog motivates and identies the main components in a
reference architecture.

Monday, August 1, 2011

Things you should know before implementing a data warehouse

"Data warehouse implementation is by nature, a complex undertaking and the degree of complexity increases depending on the size of the organisation, its politics, the stakeholder degree of buy-in and the number of data sources."

Ten Critical Factors

Wednesday, June 22, 2011

Background

Business Process (BP) "consists of a set of activities that are performed in
coordination in an organizational and technical environment. These activities
jointly realize a business goal."

Business Process Intelligence (BPI) is an upcoming buzzword in the enterprise
application area as a kind of Business Intelligence 2.0. Business Intelligence in
the traditional sense covers the spotting, extraction and analyzing of business
data. BI aims to support decision support for the operative, tactical and strategic
focus. As contrast BPI focuses on the business process data of a enterprise.
It manages "process execution quality by providing [...] analysis, prediction,
monitoring, control and optimization"[5]. For example instead analyzing the
sales revenue by products, it focuses on the productivity of the sale process
itself.

Key Performance Indicators (KPIs) are quantitative measures to capture the
performance of a business process. Each KPI is linked to a strategic goal. A
strategic goal hereby is a quanti able, measurable and result-oriented high-level
goal. By de ning av KPI one can measure if a aspect of the process serves the
overall strategic goal.

Business Activity Monitoring (BAM) is the part of BPI and aims to the real-
time monitoring of business processes to get visibility for business performance
(e.g. KPIs). Consequently, process owner and process participants get able to
react on business changes fast and more e cient.

Workflow Management Systems (WFMSs) manage and control business pro-
cesses. Consequently, they automate business tasks, mange process-related in-
formations and integrate information in the enterprise.

Audit Trail Data in general is a sequence of steps generated by a system to
document the real processing. In the WFMS, it is a set of event logs produced.
A event captured in logs represents a activity state change during the execution
of a process. This paper introduce a reference architecture for a data warehouse
specialized on audit trails logs.

Data Warehouse (DWH) is "a subject-oriented, integrated, time-variant, non-
volatile collection of data in support of management's decision-making pro-
cess". The DWH's main task is to give a uni ed view on the operational data
of a company. Thereby it o ers a variety of aggregation and analysis feature
to support the decision-making. As contrast to the traditional weekly reporting
DWs, nowadays warehouse have to face near-real time analyses requirements.
For example on going research presents the zero-latency data warehouse.

Operational Data Store (ODS) contains, as contrast to the DWH, "volatile,
current-valued, detailed-only collection of data". It's main focus is to deliver
up-to-the-second fresh information. Thereby, it enable o er near-real-time data
analytics.

Online Analytical Processing(OLAP) applications enables enterprise users to
gain insight into their data by fast access to a variety of views of data organized
to inflect the multidimensional nature of enterprise data. Thereby the data get
summarized, consolidated and synthesized according to multiple dimensions.

References:

Agrawal, R., Gupta, A., Sarawagi, S.: Modeling multidimensional databases. In:Data Engineering, 1997. Proceedings. 13th International Conference on. pp. 232{243. IEEE (1997)
Inmon, W.: What is a data warehouse? In: PRISM Tech Topic (1992)
B., I.: The operational data store. infodb. evaltech.com (1995)
Tho, M., Tjoa, A.: Zero-latency data warehousing for heterogeneous data sources and continuous data streams. In: Proc. of the Fifth Int. Conf. on Information Integration and Web-based Applications Services. Citeseer (2003)
Weske, M.: Business Process Management: Concepts, Languages, Architectures.
Springer, softcover reprint of hardcover 1st ed. 2007 edn. (11 2010), http://amazon.com/o/ASIN/3642092640/

Monday, June 6, 2011

Related Software - BizTalk Server 2009

Microsofts BizTalk server 2009 seems according to the following article to support various BAM technologies.

Microsoft delivers its own model for Business activities by introducing similar attributes as papers on activity states in common. It provides a visual interface to create execution rules for business processing and an alert mechanism which will inform the user by any fail. In addition it provides user-defined dashboards for time and costing data.

It also supports the import of KPIs to the BAM and the creation of related alerts.

To sum up, it seems that Microsoft delivers a quiet good solution for BAM by intergrating it with .NET applications and other Microsoft products.

Currently, I am not sure if one can extend it easily to interact with established process engines. Neither I am sure about the underlying performance. They divide data by running and past activities for analyses and query them by combining the data at query-time.

http://msdn.microsoft.com/en-us/library/dd320628(v=bts.10).aspx