Our world simply can not do without a lot of data. They are transmitted between different objects, and if this does not happen, then this means only one thing - human civilization has ceased to exist. Therefore, let's look at what a data stream is, how it can be controlled, where it is stored, what its volumes are, and much more.
Introductory information
First of all, we need to deal with terminology. A data stream is the targeted movement of certain information. The final destination can be the broad masses of the population (TV), electronic computers (Internet), relay (radio) and so on. There are different types of data streams. Their classification can be carried out on the basis of the means used (telephone, Internet, radio communications), places of use (company, meeting of people), and purpose (civilian, military). If their hierarchy, functional processes, related elements are of interest, then a data flow diagram (DPD) is built. It is necessary to track movements, as well as to demonstrate that each process, upon receipt of certain input information, provides logical output results. To display this position, one can construct notations corresponding to the Hein-Serson and Yordon de Marco methods. In general, the data flow model of the DPD allows you to deal with external entities, systems and their elements, processes, drives and flows. Its accuracy depends on how reliable the available source information is. For if it is not true, then even the most advanced methods will not be able to help.
About sizes and directions
Information data streams can be of different scales. It depends on many factors. For example, take a regular letter. If you write the most ordinary phrase: “Today is a good and sunny day,” then it doesn’t take up much space. But if you encode it into binary code that the computer understands, then it will obviously take more than one line. Why? For us, the phrase “today is a good and sunny day” is encoded in an understandable and non-questionable form. But the computer cannot perceive it. It reacts only to a certain sequence of electronic signals, each of which corresponds to zero or one. That is, it is impossible for a computer to perceive this information unless it is transformed into a form that it understands. Since the minimum value with which it operates is an eight-bit bit, the encoded data will look like this: 0000000 00000001 00000010 00000011 ... And these are only the first four characters that conditionally mean “this.” Therefore, processing a data stream for him is, although possible, but a specific occupation. And if people communicated in this way, it is not difficult to imagine how enormous our texts would be! But there is a downside: a smaller size. What is meant by this?
The fact is that computers, despite the fact that they, at first glance, work inefficiently, take very little space for all changes. So, to change certain information you only need to work purposefully with electrons. And the contents of the equipment will depend on where they are located. Due to its small size, despite the apparent inefficiency, a computer can contain much more information than a sheet or a book commensurate with a hard disk. Thousands, or even millions of times! And the amount of data flow that it can pass through itself grows to staggering values. So, to simply write all the binary operations carried out by one powerful server in a second, an ordinary person can take years. But there can be high-quality graphical emulation, a lot of records of changes on the exchange and a lot of other information.
About storage
It's clear that everything is not limited to data streams. They go from their sources to recipients who can simply familiarize themselves with them or even save them. If we talk about people, then we are trying to save the important in our memory for reproduction in the future. Although this does not always work out, and something undesirable can be remembered.
In computer networks, a database comes to the rescue here. The flow of information transmitted over the channel is usually processed by the control system, which decides what and where to record in accordance with the instructions received. Such a system, as a rule, is an order of magnitude more reliable than the human brain, and allows you to fit a lot of content that is easily accessible at any time. But here you can not do without problems. First of all, one should not forget about the human factor: someone missed the safety briefing, the system administrator did not take his duties with due zeal, and that’s all - the system failed. But there may be a banal error in the data stream: there is no necessary node, the gateway does not work, the format and encoding of data transfer are incorrect, and many others. Even an elementary failure of information technology is possible. For example, a threshold has been set that for nine million operations performed by a computer, there should be no more than one execution error. In practice, their frequency is much lower, perhaps even reaching a value of one in billions, but, nevertheless, they are still there.
Analysis
Typically, data streams do not exist on their own. Someone is interested in their existence. And not just in one fact that they were, but also managing them. But this, as a rule, is not made possible without previous analysis. And for a full study of the current situation, just studying the current situation may not be enough. Therefore, the entire system is usually analyzed, not just one thread. That is, individual elements, their groups (modules, blocks), the relationship between them and so on. Although the analysis of the data flow is an integral part, it is not carried out separately because the results obtained are too divorced from the whole picture. In this case, a rearrangement of entities is often carried out: some external ones are considered as part of the system, and a number of internal ones are taken out of the scope of interest. Moreover, the study has a progressive character. That is, it is first considered by the whole system, then it divides it into its constituent parts, and only then is the definition of the data flows that you have to deal with. After everything has been carefully analyzed, you can solve issues with management: where, what, in what quantity will go. But this is a whole science.
What is data flow control?
In essence, this is an opportunity to send them to specific recipients. If we talk about individuals, then everything is very simple: the information that we have is managed by us. That is, we decide what to say and what to keep quiet about.
Managing data flow from a computer technology perspective is not so easy. Why? In order to communicate certain information to another person, it is enough to open your mouth and strain the vocal cords. But this is not available to technology. Here, managing data flow is difficult.
Let us recall the already mentioned ordinary phrase: "Today is a good and sunny day." It all starts with its translation into binary code. Then you need to establish communication with the router, router, connector, or other device aimed at the received data. Available information must be encoded in order to take the form that can be transmitted. For example, if you plan to send a file on the World Wide Web from Belarus to Poland, then it is divided into packets, which are then sent. Moreover, there is not only our data, but also many others. After all, the means of transfer and transmission cables are always the same. The network of data streams hiding the world allows you to receive information from anywhere in the world (if you have the necessary funds). Managing such an array is problematic. But if we are talking about one enterprise or provider, then this is completely different. But in such cases, control usually only means where to direct the flows, and whether they should be skipped at all.
Modeling
Talking about how the data transfer stream works in theory is not a big deal. But not everyone can understand what it is. Therefore, let's look at an example and simulate possible scenarios.
Suppose that there is a certain enterprise in which data streams exist. They are of most interest to us, but first you need to deal with the system. First of all, we should recall external entities. They are tangible objects or individuals that act as sources or receivers of information. An example is the warehouse, customers, suppliers, staff, customers. If a particular object or system is defined as an external entity, then this indicates that they are outside the analyzed system. As previously mentioned, in the process of studying some of them can be transported inward and vice versa. In the general scheme, it can be depicted as a square. If a model of a complex system is being constructed, then it can be represented in the most generalized form or decomposed into a number of modules. Their module serves for identification. When posting reference information, it is better to limit oneself to the name, determination criteria, additions, and incoming elements. Processes are also highlighted. Their work is carried out on the basis of incoming data supplied by streams. In physical reality, this can be represented as processing the received documentation, accepting orders for execution, receiving new design developments with their subsequent implementation. All received data should be used to start a specific process (production, control, adjustment).
So what's next?
Numbering is used for identification. Thanks to it, you can find out which thread, where, why and how it got and started a certain process. Sometimes information fulfills its role, after which it is destroyed. But this is not always the case. Often it is sent for storage to a data drive. By this is meant an abstract device suitable for storing information that can be retrieved at any time. Its more advanced version is identified as a database. The information stored in it must correspond to the accepted model. The data stream is assigned the definition of information that will be transmitted through a specific connection from the source to the receiver (receiver). In physical reality, it can be represented in the form of electronic signals transmitted via cables, letters sent by mail, flash drives, laser disks. When constructing a schematic drawing, the arrow symbol is used to indicate the direction of the data flow. If they go both ways, then you can simply draw a line. Or, use the arrows to indicate that data is transferred between objects.
Model building
The main goal pursued is to describe the system in an understandable and clear language, paying attention to all levels of detail, including when dividing the system into parts taking into account the relationship between different components. In this case, the following recommendations are provided:
- Place at least three and no more than seven streams on each part. Such an upper limit has been established due to restrictions on the possibility of simultaneous perception by one person. After all, if a complex system with a large number of connections is considered, then it will be difficult to navigate in it. The lower limit is established on the basis of common sense. For to carry out detailing on which only one data stream will be depicted is irrational.
- Do not clutter up the circuit space with elements that are not essential for a given level.
- The decomposition of flows should be carried out together with the processes. These works should go simultaneously, but not in the order of priority.
- For designation, clear, meaningful names should be highlighted. Abbreviations are preferably not used.
When studying flows, it should be remembered that it is possible to deal with everything inadvertently, but it is better to do everything carefully and in the best possible way. Indeed, even if the person who makes up the model understands everything, then he does it, almost certainly, not for himself, but for other people. And if the head of the enterprise cannot understand what is at stake, then all the work will be in vain.
Specific Modeling Moments
If a complex system is created (that is, one in which there are ten or more external entities), then it will not be superfluous to create a hierarchy of context diagrams. In this case, not the most important data stream should be placed at the top. What then?
Subsystems in which there are data streams are better suited, and also indicate the connections between them. After the model is created, it must be verified. Or in other words - check for completeness and consistency. So, in the completed model, all objects (subsystems, data streams, processes) must be detailed and described in detail. If elements were identified for which these actions were not performed, then you must return to the previous development steps and fix the problem.
Consistent models should ensure the safety of information. In other words, all incoming data is read and then written. That is, when the situation at the enterprise is simulated and if something remains unaccounted for, then this indicates that the work was done poorly. Therefore, in order not to experience such disappointments, substantial attention must be paid to preparation. Before work, it is necessary to take into account the structure of the object being studied, the specifics of the data transmitted in the streams, and much more. In other words, a conceptual data model should be built. In such cases, relationships between entities are distinguished and their characteristics are determined. At the same time, if something was taken as a basis, this does not mean that it is necessary to grab hold of it. When a need arises, the conceptual data model can be clarified. After all, the main goal pursued is to understand the data flows, to establish what and how, and not to draw a beautiful picture and be proud of yourself.
Conclusion
Of course, this topic is very interesting. At the same time, it is very voluminous. For its full consideration, one article is not enough. After all, if we talk about data streams, then the matter is not limited to simple transfer of information between computer systems and within the framework of human communication. Here you can highlight many interesting areas. Take, for example, neural networks. Inside them there are a large number of different data streams that are very difficult for us to observe. They are trained, compared, converted at their discretion. Another related topic to remember is Big Data. After all, they are formed by receiving various streams of information about many things. For example, a social network tracks a person’s affection, which he puts a “like” mark to form a list of his preferences and offer more effective advertising. Or recommend joining a theme group. As you can see, there are many options for using and using the resulting data streams and the information contained in them.