In the discussion about outsourcing, in particular the international outsourcing of service jobs, Uday Karmakar, professor for technology and strategy at the University of
California, is adding a new view: ”The wave which we are facing and we should learn to surf, is certainly best characterized by the expression “service industrialization”. New might be the creation and utilization of information assembly
lines. Like a manufacturing company which is depending on the control of the supply chain, a service provider achieves a competitive advantage mainly by controlling the information flow.
At this point immediate access to
decision-relevant information plays an important role. In practice there is still need for optimization in the domain of mobile phone companies, conventional telephone network providers, internet service providers, card processing
Especially in the case of mass data processing even elementary statistical results on running operations often can only be delivered with a delay of weeks. If new products are launched part of the preliminary software modules
have to be adapted extensively in order to obtain relevant results a few months later as the case may be.
RSS – Advantages compared to conventional SQL programming
The ”Realtime Statistic and Simulation System” (RSS) revolutionizes your generation and processing of statistics by the
provision of statistical information which is many times faster than the common database programming via SQL-commands.
- The common approach requires at least one SQL command to be established for one statistic. This means that a data record which is relevant to two statistics for example,
has to be read twice. If a data record is relevant to ten statistics, it has to be read ten times.
- Using RSS all data records relevant to standard statistics are read exactly one time. RSS implements the summation into different statistics in a parallel way.
- In order to rise reading rate and operation speed for the generation of statistics using the common approach, it is necessary to create and to administrate quite a
number of additional database indexes. The creation and the administration of those indexes are very extensive, and depending on the internal database design are time-consuming
either during the insert of the new data records or during their subsequent reading. In each case additional indexes are noticeable because of their increased demand for disk space.
- Using RSS only a few of fundamental indexes are created. No additional indexes are
necessary in order to establish special views on single data records for example. As a result it reduces the time needed to insert and access data records, respectively. Also the required disk space for data in the database is
- With RSS you will have the possibility to define your special statistical view even after the summation of data.
- At conventional preparation of statistics the programming of the SQL command for the allocation of summation information widely determines the final appearance of your
statistic based on these summations.
- Especially in the case of new statistics being defined this can be contradictory to the aims of such statistics. On one hand you need to have a certain idea on the
results you have to expect in order to select a reasonable view, on the other hand you want to use such statistics to gain a first glance on the relevant data itself. In such cases you may not be able to determine how the reasonable
view on your statistic has to be defined.
- In many cases the consequence for new statistics prepared via conventional approach will be that you have to drop the achieved results after the first summation and to
establish one or even more new summations resulting in corresponding database load and waiting time.
- In contrast when RSS is used the single data records are pre-summarized at first, i.e. initially a summarization towards a statistic is performed,
statistical raw data are generated. The summarized data may represent the more or less aggregated initial data depending on advance information and summarization degree which can be determined by yourself.
- The RSS dialog module starts with the aggregated summarization data. Every statistical view which is (still) possible can be realized in a matter of seconds and can be
dropped eventually in order to select a modified view and to visualize it immediately again. The same summarized data can be used for more than one statistic, i.e. with the help of RSS you are able to
customize and select your statistical view straight on the present summarization data.
With RSS the initial conversion to the internal format represents only a fundamental classification. The final classification of the data within the scope of
fundamental classification can be adjusted anytime on the concrete view on a statistic without knowledge of SQL.
- If statistics are established using the conventional approach you have to choose between two evils. Either you have to classify the data towards all kinds of statistical
views already during the initial conversion of your data, i.e. you deposit the relevant information in the converted data record, or the converted data remain as they are, i.e. statistically unclassified. Naturally it is possible to
create hybrid forms.
- In the former case the build up of your statistical data can be established directly from the respective data via SQL commands. However, the possible statistical views
will be limited because of your initial classification or you have to create very complex SQL commands which will be highly error-prone.
- However, if you prefer the second option without statistical classification you will have to perform SQL JOINs with a lot of classification tables. These JOIN-components
can also result in very complex SQL commands with corresponding error-proneness. To simplify matters the effects on the runtime shall be neglected completely in this context.
- Practically a hybrid form of these both approaches will be implemented which will result in more or less complex SQL commands depending on the concrete statistical
requirements. The required SQL commands have to be generated by SQL specialists.
- With RSS no SQL commands are required for you, you do not have any SQL knowledge at all in order to define the initial summarization, to initiate its execution and
to obtain a statistic according your statistical view.
- Within the scope of the basic classification the required classes for each statistic can be assembled within the RSS dialog system individually. Alternatively you may
refer to existing classifications. You will be guided through the RSS dialog system step by step. If necessary RSS will generate SQL commands for you automatically and will start your request.
- The last mentioned procedure within RSS only affects the selected data. The classification information will be determined primary to the time of summarization and at the
generation time of the statistics with the help of the RSS control tables respectively. RSS-sided all relevant information will be held optimally administrated in the working storage so that a maximal data throughput
is ensured even for a huge amount of data.
Description of the system
Besides the handling of the running (standard) statistics RSS covers a system for the summarization of statistically relevant information during their formation process
including the visualization of these statistics in a matter of seconds ”just in time” (< 10 seconds) on the one hand. On the other hand it includes the network-wide simulation of fee-models on the basis of already existing data and also
a module for the definition and the immediate visualization of Ad-hoc statistics. These statistics are standard statistics of the respective department. It is possible to define fee-models in which special conditions like free units,
scales of fees etc. can be considered. The achieved results can be provided to subsequent systems.
In its kernel routines RSS is programmed in COBOL, hence it is capable of being easily integrated on every common computer system
. The real-time system can be addressed even multi-parallel by calls (via user-exit for example) from an existing real-time module like a authorization system or an internet system.
Because of its universal approach and the
already existing conversion and classification algorithms which are relying on this approach, RSS is able to receive and exploit data without a lot of modification effort. Company specific customizations can be easily and quickly realized
anytime by the appropriate adoption of the external parameter tables.
The definition of simulations, of real-time, standard, and ad-hoc statistics including their corresponding fees can be conducted by the user himself. Both for the
definition of statistics and their visualization dialog applications can be used which allow also subsequent modification of the statistical raw data. Thereby every access is secured by corresponding access rights.
Unique selling propositions
- RSS is able to supply statistics many times faster than the conventional database programming via SQL commands.
- RSS offers the facility to specify your statistical view even after the summarization of the single data.
- With RSS you do not have to possess any SQL command in order to define statistics, to initiate the execution of the summarizations, and to obtain a statistic
according your statistical view.
- Real-time statistics defined by RSS can be displayed with actual content on the screen
in less than 10 seconds.
- With RSS statistical and summarization information can be supplied network-wide. The aggregation and visualization can be realized multi-parallel.
- For this reason the waiting for statistical results is a thing of the past even for highest transaction rates of almost arbitrary dimension. Simulation results
as well as standard statistics can be provided in best time depending on the concrete network and process design.
- RSS is designed for day-long operation at 7 days per week.
- With RSS simulations, statistics, and fees can be defined and modified by the user himself.
- RSS contains an access rights administration and is multi-client-enabled.
- With RSS the going live of simulations, statistics, and fees during the running operation can be achieved without downtime.
- RSS is platform independent and compatible to all common database systems.
The system is fully available in banking and card-processing environments. The system performance can be demonstrated anytime within a test system.
simulation load tests with test data have been performed which can be verified and repeated anytime.