COMPUTER NETWORKS AND DISTRIBUTED SYSTEMS IN EDUCATIONAL FIELDS
COMPUTER NETWORKS AND DISTRIBUTED SYSTEMS IN EDUCATIONAL FIELDS
Rahima Jalilova
Associate Professor, Doctor of Philosophy in Mathematics "Computational Mathematics and Cybernetics Department of" Azerbaijan State Pedagogical University,
Azerbaijan, Baku
Depending on the principles of constructing the local area network (LAN) are subdivided into a "client-server" and "file server" and "peer". LAN can be composed of a means to access the distributed and global computer networks.
Client-server
Architecture and organization of building a network in which the separation is made between the computational load the inclusion of electronic computer, serve as "clients," and one powerful central electronic computer - For example, the server can maintain a central database located on a large computer reserved for this purpose. the "server." In particular, the observation of the data separated from the programs using the data. Customer will be a normal program, located on any electronic computer connected to the network, as well as the electronic computer itself, which, as appropriate, requests data from the server. Performance when using client-server architecture is higher than normal, since both the client and the server share the load of data processing. Other advantages of client-server architecture are the large amount of memory and its suitability for solving diverse problems, connectivity, a large number of workstations, including PC and passive terminals, as well as the installation of protection against unauthorized access.
File-Server
Architecture of LAN based on the use of so-called file server - a relatively powerful electronic computer that controls the creation, maintenance, and use of common information resources of the local network, including access to its database (DB) and individual files, as well as their protection. For support and maintenance of large and very large databases containing tens of millions of records, so-called multi-processor systems that can effectively handle large volumes of information and has good performance ratio price / performance. Unlike client-server architecture, this principle of construction of the network suggests that the incorporation of the workstations are full-fledged computers fitted with the full scope of work necessary for the independent composition of the core and application software. In other words, in this case no possibility of sharing the computational load between the server and the network terminals that are typical for the type of architecture of the file server, and as a consequence, the overall cost parameters of the price / performance of the network as a whole can be lower. A common shortcoming of earlier development of software tools is the fact that they were focused only on the file-server architecture for constructing computer network.
Peer LAN
"Serverless" organization building a network that allows for the inclusion of a computer as a different power, and input-output terminals. The term "ad-hoc network" means that all terminals network have the same rights in it. Each user can define peer network structure of files that it provides for general use. Thus, an ad hoc network users can work with all of your files and the files provided by its other users. Connecting individual computers in the ad hoc network is made predominantly high frequency coaxial cable links. There are three main types of ad-hoc network topologies, which are names of "tire", "ring" and "star." Create an ad hoc network provides along with the interchange of data between computers included in its share of the space, as well as the joint operation of peripheral devices. Other possibilities exist, such as when one of the computer temporarily assumes the function of "servers", and the others operate in a "clients". The latter is widely used in various types of training systems. The advantages of peer LANs are also the relative simplicity of their installation and operation, reasonable cost, the possibility of development, independence and perform computational processes to each other in a network-enabled computers. A special place in the information society takes the creation of computer networks and building on their basis of distributed systems (DS), is a set of geographically distant from each other nodes combined data system and interacting via messaging. One embodiment of the RS information processing in real life was the technology of distributed database (DDB). Each application program involves the interaction with DDB, contains one or more elementary actions (reading, writing, etc.), transforming it from one state to another. As a result of some of these actions may temporarily disrupt the integrity of the distributed database. For example, when transferring a certain amount of money from one bank account to another, at some time the first bill will be less than this amount, while the second bill has not yet received it, and thus, the total amount of money in the bank will be incorrect. To prevent such undesirable consequences of action on the data DDB grouped into transactions ─ indivisible operations, which are the logical unit of work system and its transfer from one consistent state to another consistent state. Important issues related to the organization of distributed database using transaction mechanism of simultaneous access of multiple users to the same data. These problems can be summarized briefly as loss changes uncommitted changes and a number of other, more complex problems. Additional and very significant problems may arise in the event of failures of various asynchronous nodes (processes). In DDB data is physically distributed among the nodes, each of which has some local database containing a piece of data from a total of distributed database. Therefore in distributed database transaction whose execution is to update the data on multiple nodes of the network, called the distributed transaction. A distributed transaction includes multiple local transaction, each of which terminates in two ways - fixed or liquidated. A distributed transaction is recorded only when all fixed sub transaction and its components. If at least one of the sub-transaction has been eliminated, the partial results should be annulled, and distributed transaction will be returned to its original state. Thus, following the principle of "all or nothing", observed the atomicity of a distributed transaction integrity is maintained and DDB. To implement this mechanism has been proposed a two-phase commit protocol, the main idea of which is that the transaction is recorded in two phases. In the first phase, the nodes involved in fixing, produce agreement on a common decision. The aim of the second phase - the implementation of the decision. However, a basic two-phase commit protocol and further modifications in varying degrees to the critical node failures and provides terminated in transactions. The main task of the control systems of distributed databases is to provide a means of integration of local databases, which are located in some of the computing nodes of the network, so that the user is working at any node in the network has access to all of these databases as a single database.
This should be ensured:
• ease of use of the system;
• wireless operation for violations of network connectivity or administrative needs;
• a high degree of efficiency.
Possible homogeneous and heterogeneous distributed databases. In the homogeneous case, each local database is managed by the same database. In a heterogeneous system, the local databases may apply even to different data models. Network integration of heterogeneous databases - it's relevant, but a very complex problem. Many of the solutions are known on a theoretical level, but still can not cope with the main problem - lack of efficiency of the integrated systems.
3. Object-oriented Database Management System ORACLE
Oracle was the first commercial relational database management system that supports the now industry standard language SQL; its first version appeared in 1979. In fact, all the while Oracle is the undisputed leader in the market of commercial DBMS and the second (after Microsoft) largest company producing software. Early versions of the database have been designed for the mainframe, and as the jobs use the "dumb" terminals. However, over time version of Oracle, for use in architecture "client-server" (the first version was a Oracle 5, released in 1985). Initially, these versions were designed for different server platforms different versions UNIX, VMS, and others. Later versions were released Oracle Server for Novell NetWare. The first versions of this server for personal computers emerged in the mid-90s (Personal Oracle 7 for Windows 3.1, Personal Oracle 7 for Windows 95, Personal Oracle Lite, Oracle Workgroup Server 7 for Windows NT). Before the advent of personal computers these versions can be used solely as a client workstation in the Oracle server platforms typically includes the client part for DOS. Note that Oracle was the first company that created the database, use the provided some server platforms means of parallel computing Oracle Parallel Server (before his appearance parallel computations were only used for scientific problem solving). When using parallel computing Oracle Parallel Server enables multiple processors to access the same database, which allows for high-speed transaction processing, and later versions allow it to decompose operations with large volumes of data in order to perform their parallel on multiple processors.
To date, the latest version of Oracle version is Oracle 8i, the distinctive features of which are:
• availability of object extensions and corresponding data types such as nested tables, arrays, objects, and others. In other words, Oracle 8 and Oracle 8i are object-oriented database management system;
• Availability of analytical data (for example, calculating percentages, rankings, comparisons of time periods);
• the ability to create tables containing aggregate data (materialized views) and the possibility of partial update them when the data on which they are calculated;
• Support for Java, in particular JDK 1.2 and JDBC 2.0;
• Support for XML, in particular in the Oracle 8i included XML Parser for Java, C / C ++, PL / SQL, XML-transforming the data into a form suitable for use in Oracle 8;
• support HTML- and XML-pages incorporated with the code PL / SQL (for their implementation, additional products, such as Web DB PL / SQL Gateway or Oracle Application Server PL / SQL Cartridge);
• Support for storing multimedia data with the possibility of indexing context of constructing queries, support for different languages for stored documents;
• a set of procedures and functions for processing spatial information (Oracle Spatial);
• additional security features, such as data encryption, support for SSL, use database-level roles and the enterprise;
• establishing systems, fault tolerance by using multiple parallel processes;
• Support for Microsoft Cluster Server;
• Availability of OLE DB-provider to access data.
Oracle 8i exists in three editions: Oracle 8i, Oracle 8i Enterprise Edition, Oracle 8i Personal Edition. To create a multidimensional data warehouse exists a separate product Oracle Express OLAP. In addition to the different versions of the database server of Oracle products is also available Designer / 2000 focused on the DBMS CASE-tool for business process analysis and design data, and funds development of client applications.
One Developer / 2000 (previously called Oracle * Forms) is very popular among users of Oracle; there were other development tools (for example, Oracle Power Objects). Note that applications built using Developer / 2000 can run on different platforms. Language PL / SQL, used in this development tool is interpreted and is the same language that is used in Oracle for writing server code. This allows you to debug using Developer / 2000 server code. By producing their own development tools, Oracle provides its users the ability to create client applications by other means. In particular, in addition to the standard in such cases, the client API (Oracle Call Interface) client part of Oracle also contains the object model (Oracle Objects for OLE), allowing the use of the client part of Oracle as a set of COM-objects to access data. In addition, usually the client part of Oracle also contains ODBC-driver to access the data of the database. Note that many other companies produce ODBC-drivers and OLE DB-providers for access to Oracle (in particular, Microsoft). Companies producing development tools that use native libraries to access data (such as In prise or Gupta / Centura), also includes a library of access to Oracle in the most expensive versions of their products. From ready information systems based on Oracle particular note are several large enterprise management systems, in particular SAP / R3. The West is also often used ready-made solutions from the most Oracle Corporation, united under the common name of Oracle Applications, such as Oracle Financials, Oracle Human Resources, Oracle Market Management, Oracle Project Systems, and others.
References:
- Brown, S. "Mosaic" and "World Wide Web" to access the Internet: Per. from English. - M .: Mir: Malip: SK Press, 1996. -167 p.
- Gaykovich V. Pershin A. Security of electronic banking systems. - M .: "United Europe", 1994. -264 p.
- Levin, VK Protection of information in computer systems and networks // programming. -1994. -N5. -C. 5-16.
- AV Frolov, Frolov GV Global networks of computers. Practical introduction to the Internet, E-mail, FTP, WWW, and HTML, programming for Windows Sockets. -Dialog -MIFI, 1996. -283 p.