We live in real time, minute by minute. News did not come with a delay of days, they are transmitted in the air. Banking operations are carried out expeditiously, the account balance can be determined at the moment. You can book flights in real time, assessing which sites are free. All these operations generate a lot of data.In
order to adapt the business model for the modern world, software tools
designed to help technology driven by external events (event - driven). Data
is transmitted in real time through a service-oriented architecture
using loosely coupled and providing good service interactions, which
stimulate the standardized integration of applications.Since
the invention of relational databases and SQL - query technology
business intelligence (BI) did not change fundamentally, but now there
is a new concept of BI.A
new concept of BI includes several important new concepts regarding the
use of information in business organizations and government agencies. The
term is inherently related to BI in the real-time technology driven by
external events, but the basic idea is to apply these techniques to
business processes.At the heart of this architecture developments, particularly XML - messages. After
all, most modern processes are run events, and, therefore, if thinking
about how to make current processes more intelligent, ordinary SQL -
queries may seem far from ideal.Traditional
stores gave rise to a significant development in the use of
information, but now the underlying architecture is already causing some
problems. This architecture limits the possibilities for optimizing all business processes for embedding BI - opportunities. Please refer to the event-driven analysis of process, which will replace the batch reporting, done after the fact.In a nutshell: how to build flexible business processes that give organizations a competitive advantage? How to build a "smart" business?Client-server model obsoleteBI
- tools used today in most organizations, have been developed to solve
the problem that arose in the early nineties due to the development of
relational databases. And
as the bases accumulated more and more information, simply extract it
became routine for IT - department, as most users are not interested in
acquiring new skills - expert in writing SQL - queries. Retrieving data from a database has outlived its usefulness and has led to the development of BI. Therefore, existing to this day BI - tools focused on data representation.As
it turned out, extracting data corresponding to events that occurred a
few days or hours ago, and publishing them in the form of reports, has
its benefits, but do not provides users with guidance for action to
improve efficiency at a given moment. As a result, many companies BI - users do not revise their submitted reports, but simply refer them to the background. Complaints against an aging and losing the benefit information can be heard frequently.Amazingly,
the fact is that this situation - the exact opposite of effective
intelligence in real time, which is necessary for most companies to
ensure quality of service. One might say that such information is delayed for a day and not doschityvaet dollar in most industries. For
example, in retail trade, 3 - 4% of potential revenue is lost due to
malfunctioning of the warehouse: the manager requests a report, which
comes in the morning or after working hours, it's too late to replenish
stocks.Until
now, all development BI - systems was reduced to faster query
performance for data warehouse and the more successful reports and
dashboards. And it does not solve the problem because it does not affect the architectural aspect. There
is no denying the fact that at the time of entering data into the
warehouse and their subsequent recovery, they become obsolete. For some applications, there is no problem, but for those who need the information in real time - the situation is deplorable.There is a widespread misconception that real-time data are not needed, since the task forces can not analyze them. The mere provision of reports in a shorter time will not solve the problem. So what do you need? It is necessary to enable the operational divisions in time to detect differences in daily operations. Reports - this is not the main result of BI - the system. Requires analysis and interpretation before a decision is made. It is obvious that users are not looking for them until they understand that the problem is long overdue. Rather
than report on the effectiveness of the process after the fact, BI - a
means should be applied in the process as a way to control an automated
workflow based on the actions of the client. To achieve this, we must not only collect data in real time, but also to analyze and interpret them. This is the essence of BI - technology-driven events. Data
for the last minute are analyzed in the context of historical
information, so that all actions can be initiated automatically. And then the data warehouse will not solve all problems, as projected for other purposes.The emergence of BI - servicesOver
the past few years, companies have to submit their data warehouse as
the web - services for use by other applications and processes related
to service-oriented architecture or middleware, such as ESB (enterprise
service bus - ESB). The
fundamental limitation of this approach is that the repository is not
the source, in which you want to look for data to assess the
effectiveness of the current process. Needed to analyze the data on the status of the process, it is unlikely to be found in HD.Even
the use of a dashboard with data warehouse is not suitable for most
operating tasks, since it relies on the user detects the problem with
the disposal of obsolete data. Dashboards gather and averaged data. They removed the details and context, and only information about the past. Solutions require the details, and must be taken now.Obviously,
the data warehouse will remain, but his main role will be recording the
data, not the source of the implementation of BI - operations. Reporting and presentation of historical data will continue to be important and their implementation in the CD to continue. Attempts to transfer the data warehouse in real time, not very effective. Obviously,
the information needed to support decision-making should be drawn daily
from a different source to eliminate the delay associated with the
cycle of extraction, transformation, loading and querying.A new viewIf
we assume that the goal of BI is to reduce the time lag between the
events and actions taken, thus improving business efficiency, the
existing BI - Architecture hinder this goal.The infrastructure of the BI data is not stored in the database and retrieved for analysis. It
applies processing streams of events, or in parallel with the actual
business processes, or as a separate step in the process.Usually,
this means that you need to search for script events (such as patterns
and combinations of the sequence of events), it is important to resolve
problems quickly. At
the output of such systems - usually indicators and warnings, as well
as the immediate initiation of actions in other applications. As
a result of the analysis process is automated and does not require
constant human intervention, but in certain situations, a person is sent
a request to perform some action.In the BI system data comes from a middleware, which is the most convenient source of data in real time. Standard middleware easily generates a stream of events for analysis, performed in memory. When
these events are compared with those in the past, you can quickly and
automatically identify the problems and their possible resolution.Intellectual processesTo change benchmarks, companies need to achieve a more flexible processes. And
that means a higher level of automation, or to provide effective
information operations staff and changing from day to day operating
procedures, management processes. A
solution could be using technology messages (messaging technologies),
supported by transaction systems, tools, business process management and
SOA, and BI - the new generation technology. All of this gets on very well together, we can imagine a real-time BI analysis of both services in a SOA architecture.BI technology is applicable to both well-defined processes, and to and less certain areas. Many processes can be modeled and explicitly defined by the means of managing business processes. Business users often can not accurately describe their processes, but the analysis in this case is also required.A new concept of BI has a number of properties:
Event-driven. Automated processes are event-driven, so it is obvious that the development of more flexible processes to analyze and interpret events. So, to analyze the data from event to event, as in parallel with business processes, as well as a separate stage of the process.
Performed in real time. Otherwise it is impossible to implement BI - the possibility of a step in the process, and thus no longer be able to automate the operation. For comparison, batch processes - information they contain reports on the effectiveness of the process, but can not be part of the process, except in cases where time does not play a decisive role. Any application which is relevant to trade, dynamic pricing, demand estimation, safety, risk, detect fraud, replenish stock and any interaction with the client - a process that depends on time, and therefore requires processing in real time.
Automated analysis. To automate the daily process of decision-making, organizations should not simply present data in a dashboard or report. The task - to turn data into real-time into something effective. Necessary to automatically and dynamically interpreted. In practice, this implies a comparison of each individual event that is expected to be normal, based on past experience or anticipated future performance. BI products should be used as the basis of a norm as to the detailed and aggregate level and compare specific events automatically to this standard.
Foresight. To understand how this event will affect the needs of your organization, you must have a vision. To answer the questions: "Will the order on time?", "Deny whether the system today?" - Should be able to make predictions. This feature emphasizes the operational divisions, which should be thought of the prospect of changing the effectiveness of their work during the day, week or month.
Focusing on processes. BI products of the new wave should focus on processes. This does not mean that the processes are modeled using a management tool. Actions can be optimized based on the results of a specific process, although he may not have a precise definition.
Scalability. Scalability is the cornerstone of BI. Flow of events can be unpredictable and very intense. For example, if retail demand estimates are developing an application to track sales of the most popular products, then we have the following situation. The company can be, for example, 30 thousand products sold in a thousand stores. The result is 30 million combinations of "shop-goods" that must be monitored. The day can be sold to 10 million items. This scale, the situation is commonplace for a new concept BI. In fact, even such a scaling itself requires the application of a new class, which in traditional BI technology was not available.BI - the technology of real time, event-drivenBI represents a fundamental shift to a different use of the information. This technology is the development of BI and goes beyond the traditional data warehouse and query tools, including dynamic and automated decision making.In the past, companies were forced to rely on outdated information and try to solve the problem is when they arise. A new concept of BI makes a difference because it allows you to embed analytic capabilities in the processes themselves, that is to make the latter more flexible.Now, problems are identified and immediately take corrective action. And this is a step towards providing truly useful information that will change the performance in general. Do not expect this from BI - technology from the very moment of its appearance?
Event-driven. Automated processes are event-driven, so it is obvious that the development of more flexible processes to analyze and interpret events. So, to analyze the data from event to event, as in parallel with business processes, as well as a separate stage of the process.
Performed in real time. Otherwise it is impossible to implement BI - the possibility of a step in the process, and thus no longer be able to automate the operation. For comparison, batch processes - information they contain reports on the effectiveness of the process, but can not be part of the process, except in cases where time does not play a decisive role. Any application which is relevant to trade, dynamic pricing, demand estimation, safety, risk, detect fraud, replenish stock and any interaction with the client - a process that depends on time, and therefore requires processing in real time.
Automated analysis. To automate the daily process of decision-making, organizations should not simply present data in a dashboard or report. The task - to turn data into real-time into something effective. Necessary to automatically and dynamically interpreted. In practice, this implies a comparison of each individual event that is expected to be normal, based on past experience or anticipated future performance. BI products should be used as the basis of a norm as to the detailed and aggregate level and compare specific events automatically to this standard.
Foresight. To understand how this event will affect the needs of your organization, you must have a vision. To answer the questions: "Will the order on time?", "Deny whether the system today?" - Should be able to make predictions. This feature emphasizes the operational divisions, which should be thought of the prospect of changing the effectiveness of their work during the day, week or month.
Focusing on processes. BI products of the new wave should focus on processes. This does not mean that the processes are modeled using a management tool. Actions can be optimized based on the results of a specific process, although he may not have a precise definition.
Scalability. Scalability is the cornerstone of BI. Flow of events can be unpredictable and very intense. For example, if retail demand estimates are developing an application to track sales of the most popular products, then we have the following situation. The company can be, for example, 30 thousand products sold in a thousand stores. The result is 30 million combinations of "shop-goods" that must be monitored. The day can be sold to 10 million items. This scale, the situation is commonplace for a new concept BI. In fact, even such a scaling itself requires the application of a new class, which in traditional BI technology was not available.BI - the technology of real time, event-drivenBI represents a fundamental shift to a different use of the information. This technology is the development of BI and goes beyond the traditional data warehouse and query tools, including dynamic and automated decision making.In the past, companies were forced to rely on outdated information and try to solve the problem is when they arise. A new concept of BI makes a difference because it allows you to embed analytic capabilities in the processes themselves, that is to make the latter more flexible.Now, problems are identified and immediately take corrective action. And this is a step towards providing truly useful information that will change the performance in general. Do not expect this from BI - technology from the very moment of its appearance?
