Common databases. Database management systems. Databases for your data

Popular corporate databases

Introduction

It would be completely wrong to say that the market databases is still immature. Many of the database management systems mentioned in our list date back to the mid to late 70s of the last century.

These corporate ones are still very competitive to this day and are literally crammed with functionality that ensures their high availability. The price of these DBMS varies from zero to tens of thousands of US dollars.

In the world of data management, there is no single universal solution, so arguing about which of the existing DBMS is the best does not make sense. Each of them has both advantages and disadvantages.

If you need a database application, there are at least the top 10 most popular DBMS solutions on the market today.

Decide for yourself which one best suits your requirements.

1.Oracle

Oracle traces its lineage back to 1979. Historically, it is the first commercial relational database management system (RDBMS). Since then, the Oracle name has become synonymous with enterprise RDBMS, reliable data delivery, and fierce market competition.

High-tech, sophisticated solutions have been and remain the backbone of this company, which is included in the ranking of the 500 largest companies in the world according to the Fortune business magazine (current position 105 and 27 in terms of profitability).

2. SQL Server

Whatever they say about Microsoft and its management, their profitability leaves all IT companies far behind, and a product like SQL Server is far from the last place here.

Desktop systems from Microsoft are ubiquitous, but if you suddenly need Microsoft Server, then, with a high degree of probability, SQL Server will also be running on it.

SQL Server is easy to use and tightly integrated into the operating system, making it a natural choice for Microsoft-based enterprise infrastructure.

Microsoft is currently promoting its SQL Server 2014 as a big data processing and business intelligence platform.

3.DB2

The Blue Giant (IBM) once again justified its name by launching data centers with DB2 installed on them. DB2 has been ported to Linux, UNIX, Windows, and mainframes. With the release of DB2 9.7, IBM challenged Oracle 11g because, other things being equal, it promised significant savings when migrating from an Oracle brainchild. How significant? Approximately 34-39% over a three-year period of operation.

4.Sybase

Sybase is still a serious player in the enterprise database market after 25 years of successful promotion of its Adaptive Server Enterprise. Although its market share has dwindled significantly in recent years, Sybase still has a strong foothold in extreme transaction processing.

In addition, Sybase is a partner solutions provider for the mobile and embedded markets.

5. MySQL

MySQL was originally a niche RDBMS for developers, but has since grown into a powerful competitor in the enterprise database market. MySQL is currently part of the Oracle empire (January 2010). No longer solely a niche product, MySQL supports hundreds of thousands of corporate websites and a huge number of commercial applications.

Although the MySQL community had some doubts about Oracle's right to own this free software, Oracle has publicly confirmed its intention to continue to develop and support this product.

6. PostgreSQL

PostgreSQL is the most advanced open source RDBMS. This database is invisibly present in online gaming applications, data processing automation applications, and domain registrar servers. PostgreSQL finds its way into Skype, Yahoo!, MySpace, and can often be found in the most unusual and unexpected places.

7. Teradata

Have you ever heard of Teradata? If you have ever organized a large corporate data warehouse, then for sure. In the 70s of the last century, Teradata laid the foundations for building the first data warehouses. In 1992, they first created such a warehouse for Wall-Mart, and since that time, the concepts of "corporate data warehouses" and Teradata have become inseparable.

8.Informix

Another IBM product that comes in several versions at once. Informix is ​​the world's number one college and university database system for meeting user expectations. Informix users often emphasize its low cost, low operating costs, and high level of reliability.

9. Ingres

Ingres is the parent project for PostgreSQL and other open source RDBMSs. Using Ingres lowers the total cost of ownership of the system. In addition to its attractive business model, Ingres includes tools that make it easy to migrate from more expensive databases. Ingres also includes fully standards-compliant security tools.

10. Amazon SimpleDB

Although it may seem that the concepts of databases and Amazon are not very compatible, but in fact they are not. Amazon's SimpleDB offers the enterprise sector a simple, flexible, and low-cost alternative to other commercial counterparts.

SimpleDB is characterized by low operating costs, speed, high scalability and tight integration with Amazon services.

If you're planning to get involved in any kind of serious development, knowing databases is vital. They are everywhere, even if you don't know it. The PYPL ranking, which calculates the popularity of databases, identifies 14 database management systems that are used by at least 1% of users. PYPL indexes popularity based on the number of requests in Google search. As of October 2017, Oracle remains the most popular database, with a score of 32.26%, according to the processing of the collective intelligence results. This cross-platform DBMS is the elder of the database world. It's hard to believe, but the first version, which for reasons unknown to us is called Oracle v2, was released back in 1979. The newest to date - Oracle 12.2 - in September 2016. The Oracle database is most commonly used by large companies, but there is also a free version if you want to practice so you can then find a job related to the "adult" version. It's called Oracle Database Express Edition and can be downloaded from the Oracle website. Of course she has a large number of restrictions (it uses up to 1 GB of RAM, only 1 processor, only one DB up to 11 GB).

Second place went to MySQL. Those who studied databases at university or tried to master them on their own, most likely used this free database. This is an interesting solution for medium and small applications, now also owned by Oracle. By the way, MySQL passed into the possession of this company along with our favorite language Java, in the process of acquiring Sun Microsystems. According to PYPL, MySQL is of interest to 21.15% of the total number of users or programmers.

In third place with comfort and 16.78% is Microsoft's SQL Server. This DBMS is great for managing databases of various sizes, from personal to Enterprise. It has also been on the market for a very long time, albeit a decade less than Oracle: April 24, 1989 is considered to be its date of birth. Like the vast majority of Microsoft products, SQL Server is paid (in the case of the Enterprise version, even very paid), but there are also free options. The simplest of these is SQL Server Express, which is well suited for educational purposes and can be deployed on small servers or regular computers. Database size - up to 10 GB. There is also a free version for programmers. It is intended for creating, testing and demonstrating an application based on a "real" SQL Server. The Microsoft product is followed by PostgreSQL by a wide margin. This open-source development of the University of California at Berkeley has long been in the shadow of its competitors, but in recent years its share has been slowly but surely growing. And all thanks to its free, easy extensibility and a system of built-in programming languages.

Fifth and sixth places went to non-relational databases MongoDB and a very young Firebase. Interest in both is clearly growing, especially in Firebase: compared to October last year, interest in it grew by 1.4% and now its result is 2.19%. Rounding out the top ten are SQLite, elasticsearch, Apache Hive and DB2.

PYPL rating of DBMS popularity, October 2017 and growth compared to October 2016

1. What trends in the development of server DBMS could you note in 2015–2016?

Vitaly Chesnokov, QSOFT
The most important trends in the development of modern DBMS: the use of virtualization and GRID technology, self-diagnostics and automatic correction, the use of NoSQL DBMS in big data, using NewSQL DBMS, executing C/C++ code in the DBMS address space.

Over the past few years, the amount of data suitable for processing and storing in a database has grown exponentially. An amendment to the law “On Personal Data” was adopted, stating that personal data of citizens of the Russian Federation must be stored on the territory of the Russian Federation. Some Western countries also have similar laws. All this leads us to the need for clustering and splitting data into parts.

Everywhere, the percentage of using NoSQL DBMS is growing, where possible, due to the high speed of working with data and the possibility of relatively simple clustering. A new type of DBMS, NewSQL, is gaining ground. The main unprecedented features of NewSQL include: the possibility of asynchronous master-master replication, which replaces the classic master-slave scheme and provides greater flexibility for high-load projects; simplification of administration and provision of dynamic database management; support for C/C++ stored procedures and the ability to execute C/C++ code in the DBMS address space (provide virtually unlimited extensibility and an incredible performance boost); improved diagnostic and debugging tools.

In addition, the use of virtualization in the DBMS provides the necessary fault tolerance and scalability.

Nikolai Fetyukhin.MST
Transition to NoSQL and specialization of databases. For example, you can pay attention to Redis and Tarantool. The latter even contains its own application server. An interesting trend is the combined DBMS and backend, like Facebook's Parse. Also smooth migration of databases to the clouds.

Petr Urvaev SimbirSoft
Functions that have proven themselves successfully in some DBMS are implemented in other products after some time. For example, materialized views, which first appeared in Oracle DBMS, were later implemented in MS SQL Server, and then appeared in PostgreSQL. The benefits that NoSQL solutions provide are gradually being implemented in relational DBMS as well. For example, in latest versions PostgreSQL supports working with data in JSON format.

Evgeny Gusev, ITECH
Changes in recent years in the DBMS segment were both private - in relation to individual leading products, and structural in nature, so there are many trends. First, heterogeneity. The transition to the microservices model made it possible to flexibly select the means of solving the data storage problem, not being limited to one. Secondly, the development of NoSQL, in-memory storages. Thirdly, Big Data is a revolution that required rethinking both the methodology for storing data and the very concept of “data”. Fourth, columnar (column-oriented) databases.

2. In your opinion, is there a tendency for DBMS to move to the "clouds"? What are the pros and cons of this approach?

Vitaly Chesnokov, QSOFT
Yes, there is definitely a trend. To begin with, we need to separate two fundamental approaches to the operation of a DBMS in the cloud.

The first is deploying a virtual machine with a DBMS in the cloud. You can upload your own image to it or use a pre-prepared one with an already optimized DBMS. In fact, such a virtual machine is not fundamentally different from a regular physical server. The main advantage compared to a physical server is the ease of scaling, both vertically (you can allocate more resources for this "virtual" at any time) and horizontally (creating a new "virtual" takes only a few minutes). Another significant plus is the high availability of cloud virtual machines(99.9%–99.99%). Cloud hosts also provide many additional services such as monitoring, backup, server control panel, etc.

A fundamentally different approach is a cloud DBMS. In this case, the client does not buy a server, but simply the service of using a DBMS. The current market of public cloud DBMS, amounting to $400 million, will increase to $1.2 billion by 2017. The main advantages of this approach: payment for not provided resources (which may “idle”), but only for those actually used: the amount of data stored, the amount of processed DBMS operations; there is no need to configure and administer the DBMS - these tasks are completely on the hoster; no need to think about scaling; the hoster provides many convenient and intuitive tools for managing the DBMS; high availability. The main disadvantage is the inability fine tuning DBMS.

You can also separately highlight such a subspecies of a cloud DBMS as DbaaS (Database as a Service). Almost always, a specific DbaaS is one specific DBMS provided in the cloud by direct developers. From this, the difference in business models is clearly derived: cloud DBMS are suitable for large-scale typical tasks, while DbaaS is suitable for specialized ones, for a specific brand of database engine, with the possibility of direct communication with its developers. In addition, DbaaS allows you to much more accurately select the system for the right load, in particular, by regulating the number of client connections.

Nikolai Apurin, Artwell
The trend exists. Most large (all-Russian) systems already use cloud databases.

Nikolai Fetyukhin.MST
The trend is weak. The use of clouds, although it reduces some costs, but leads to new costs. Using the cloud can only be beneficial for projects with low attendance.

Advantages of clouds: easy scalability, high fault tolerance, availability of servers around the world, easy cloning and deployment of data. Cons: while the clouds, as a product, are "raw" - it is impossible to physically control the data, since they are managed by the cloud provider.

Petr Urvaev SimbirSoft
The trend of moving DBMS to the clouds exists as part of the general trend of moving the entire IT infrastructure of organizations to the clouds. The main advantage of this approach is the ability to transfer database maintenance. The disadvantages include the storage of important data on an uncontrolled site. A cloud storage user, if necessary to destroy the stored data, cannot be sure that the data has actually been erased.

Evgeny Gusev, ITECH
It exists and is one of the primary ones. There are a lot of advantages of placing a database in the cloud: ease of sharding and replication, the ability to effectively separate data from business logic in the context of performance, easy and understandable management of computing power, and a number of others. Critical minuses, based on experience, we do not see. The more data you have to handle, the more horizontal scaling is required, the more justified the use of clouds becomes.

3. What factors influence the choice of a DBMS? For which projects are SQL databases more suitable, and for which - NoSQL?

Vitaly Chesnokov, QSOFT
The main factor in choosing between SQL and NoSQL DBMS is the needs of the application. SQL is better suited for some tasks, NoSQL is better for others.

There are several key differences between these types of DBMS. The data format in SQL is very strict, there are clear table schemas indicating where which type of data is used. There are no predefined document schemas in NoSQL - any information can be added to any document.
In SQL, there are complex relationships between different tables. Data in one table is often a link to data in another (data normalization principle). In NoSQL, as a rule, each document is an isolated information unit and stores all available data (denormalization principle).
SQL has built-in mechanisms to maintain data integrity (for example, you cannot delete a record from a table if it is referenced in other tables). There are no such mechanisms in NoSQL, so data denormalization is important (ideally, each document stores absolutely all information about any object).
SQL has a transaction mechanism that allows you to execute multiple SQL queries on an all-or-nothing basis. In NoSQL, a similar mechanism exists only within a single document.
Ideally, NoSQL is faster than SQL at the cost of more simple method data storage, which allows you to get all the information about the object simple request one document. However, there is a problem associated with the fact that the most popular NoSQL DBMS appeared quite recently. The consequence of this is a smaller amount of information on these DBMS and a larger number of unresolved problems.
NoSQL is much easier to scale due to the lack of complex logical relationships between documents. In addition, most of the NoSQL DBMS was originally created with an emphasis on scaling mechanisms.

As a result, NoSQL is better suited for projects with a large amount of data that can be easily divided into separate independent objects. Provides high speed and scalability. SQL is suitable for projects where various data have complex logical relationships with each other and their integrity is extremely important.

Nikolai Apurin, Artwell
NoSQL - for non-standard calculations with a huge amount of data. But as practice has shown, volumes up to 20 million records are perfectly processed by SQL databases.

Nikolai Fetyukhin.MST
NoSQL technologies are actively used by well-known companies, including in high-load projects. Saving data and simple selects with NoSQL will be really fast. In the case of more complex queries, the task will have to be solved on the product side, which complicates the product itself. In its purest form, we do not choose NoSQL. The complication of the product logic and emulation of basic SQL things leads to an increase in the cost of the project. And not every NoSQL solution provides data security in critical situations.

Petr Urvaev SimbirSoft
The choice of a database often depends on the preferences of the architect, the possible load, the required functionality. SQL databases allow you to clearly define data storage schemes and retrieve data using complex queries, NoSQL databases allow you to store data in a less ordered format and support horizontal scaling. Often, distributed systems use both SQL and NoSQL databases, each of which solves its own problems.

Evgeny Gusev, ITECH
In the current state, SQL / NoSQL are rather not competing, but rather complementary entities. Using SQL solutions in one application, when you need to work with complex data in their relationship, and NoSQL, when the speed of working with unstructured information comes to the fore, is a completely natural practice.

4. How would you rate the distribution of paid DBMS licenses among users? When does it make sense to buy a license?

Vitaly Chesnokov, QSOFT
There are two different options for dividing the DBMS into paid and free.

First - free versions commercial DBMS (MS SQL, Oracle, etc.) In fact, this is a stripped-down version of the DBMS, which lacks some of the functionality. Here the main factor of choice is very simple - whether this functionality is needed for this project. Less often there is a free version, which does not differ from the commercial one in terms of functionality, but is updated less often (Couchbase Server).

The second is free DBMS, for which there are similar commercial products (MySQL from Oracle, Percona Server or MariaDB). In this case, the advantage of a commercial product usually lies in the availability of more serious technical support. In terms of functionality, paid and free DBMS differ little from each other, although both of them may have their own tools that are not available in other versions.

As a result, there are two main reasons to choose a paid DBMS: the presence of functionality that is not available in free counterparts, and the ability to contact the manufacturer's technical support for help.

Nikolai Apurin, Artwell
Why pay when there are free ones? However, there are many solutions that can only work with paid databases. Basically, these are foreign practices.

Nikolai Fetyukhin.MST
The question is ambiguous. Often, paid DBMS have free versions, and free ones have components that can be purchased for money. The difference is most often in the built-in analytics and database monitoring tools. Therefore, paid DBMS are more suitable for large projects with large distributed systems.

Petr Urvaev SimbirSoft
Paid DBMS licenses are preferable to free ones when the project relies on the capabilities of a particular database, and it is important for it that the features used work as advertised, and problems in their work are quickly eliminated. Now commercial and free databases are used equally often, and most new projects choose free databases, since the data handling capabilities and stability are approximately on the same level as paid and free DBMS.

The concept of a database has become so firmly established in our lives that it has become taken for granted and does not deserve special attention. About IT startups, algorithms , hacker attacks, cryptocurrency, (Yes, what can I say, and about the clouds too) in the specialized media, much more articles have been written than about “ordinary”, but such important and necessary databases and their management systems. Here we are, repeatedly conducting our mini-research of various areas of the IT world, but have never addressed this topic. Well, we are immediately correcting and sharing interesting news and fresh statistics with you.

And among the DBMS there are favorites

A database management system (DBMS) is a software tool (usually an interface between an end user/application and the database itself) that makes it easier and more convenient to work with information. For example, create, update, search, delete and restore data in the database, as well as determine the relationship between its components (tables).

Typically, a DBMS includes three main components: the data itself, the database engine, and a schema that defines the logical structure of the data. It is these three components that help ensure safe management anddatabase protection,the integrity of the information stored in it and unified administration procedures - change management, configuration and performance control, backup, disaster recovery etc.

Of course, there are incredibly many database management systems themselves, but there are hardly a dozen of those that everyone hears about. To determine which of the existing DBMS rightfully falls into the list of the most popular, DB-Engines compiled a kind of rating of favorites. To do this, experts analyzed a number of factors: the number of mentions of systems on websites (Google, Yandex and Bing search engines were used), the general interest of users in Google Trends, mentions in discussions on specialized Stack Overflow and DBA Stack Exchange sites, on Indeed job aggregator sites and Simply Hired, on Linkedin and Upwork professional profiles, and finally the number of relevant tweets. However, we note that the total number of DBMS installations was not considered, so this top turned out to be interesting, but still relative. According to DB-Engines, the top three popular management systems are: Oracle, MySQL, Microsoft SQL Server.

Interestingly, among programmers, the skills of working with such DBMS as MySQL, MongoDB and PostgreSQL turned out to be the most in demand, and developers generally put Oracle in last place, since only 12% of respondents noted this system (source: Stack Exchange (Stack Overflow Talent)) .

Thus, according to the results of two studies, we see that so far users are most interested in the MySQL system. Perhaps not least here is the fact that this is an open source product. By the way, it is for this reason that the largest number of vulnerabilities were eliminated in MySQL in 2016 - 133. And this fact does not mean at all that the system itself is poorly protected, but rather indicates that its performance was checked by a larger number of specialists, which increased chances of identifying its “weak points” in information security.

Common security threats

It is noteworthy that it is the number of patches that makes it possible to indirectly determine the degreeinformation protection in database management systemsand identify the most common vulnerabilities. Thus, a study by Trustwave showed that in 2016 DBMS users most often encountered such categories of cyber threats as:

• Unauthorized privilege escalation.These vulnerabilities allowed unauthorized individuals to use administrative rights to access tables and database configurations.
• Buffer overflow.This led to a server crash, which underminedhardware database protection,and also caused a denial of service and could lead to the execution of someone else's malicious code.
• Permissions configured by default.administrator accounts, left with the default password could provide additional scope for fraudulent activities by cybercriminals.

These are the results of recent studies. We really hope that in the future database management systems will become more secure, and new names will appear in the tops of DBMS popularity. We promise to follow the news and keep you up to date with the latest trends.

Few would argue that the IT of the future is inextricably linked to the use of huge databases. The world is already inventing new languages, new algorithms, just to simplify and speed up the use of huge flows of information. Even the relational approach familiar to many modern users is slowly but surely becoming a thing of the past. Why and what will happen next? However, let's talk about everything in order.

From past to present

It makes no sense to cover the history of databases, clinging to any similarity, so the moment of the emergence of databases will not be ancient times, but the 60s of the 20th century. It was then that computers became an effective tool for commercial companies, and the COBASYL (COnference on DAta SYstems Language) organization, which created the COBOL language in 1959 and subsequently endowed it with database management capabilities, helped them manage the dramatically increased information flows.

By the end of the 1960s, the first network model data, the concept of a DBMS arose, and in 1974 IBM began working on a language for System R. This is how SEQUEL (Structured English QUEry Language) was born. However, later, when it became known that such a name was used by a British aircraft manufacturer, it was decided to slightly reduce it to the usual SQL.

With the increasing availability of computers, consumer-oriented databases (Paradox, RBASE 5000, RIM, dBase III), APIs (ODBC, Excel, Access) and development tools (VB, Oracle Developer, PowerBuilder) began to appear. Of course, the trend has also embraced the Internet, today effective interaction with the database is an unspoken requirement for any resource with more or less dynamic information.

As far as companies are concerned, the market has become a trinity of powers, with virtually all database power shared between IBM, Microsoft, and Oracle.

Present and future

Until the start of the new millennium, IT was dominated by a relational approach to databases, but the need to improve performance inevitably led to the development of the idea of ​​​​NoSQL (not only SQL). If you can hardly imagine what it is and what is the difference, then by clicking on the link you will receive comprehensive answers to all your questions.

To put it simply, the relational approach describes data in the format of tables, that is, all information is inextricably linked by relationships and structure (remember Excel with columns and rows, where each new object is written according to the same template). This inevitably leads to performance and scaling limitations, but from the point of view of creation and management, it is simple and convenient.

The NoSQL approach avoids these problems due to the lack of strict information links. But here another problem arises - the organization of access. It is solved in 4 main ways: using document orientation, expandable records (sparse matrices), access keys and graph theory. Naturally, the NoSQL approach requires more knowledge and skills from the developer, but the results are much more effective. That is why it is believed that SQL is already going down in history, and NoSQL is the future of all databases.

However, this prediction rests on the fact that using the relational approach for small databases is much more efficient. Therefore, instead of a senseless dispute, let's talk about more practical things, namely directly about the most popular databases.

Rating

1. Oracle;
2. MySQL;
3. Microsoft SQL Server
4. PostgreSQL;
5. MongoDB;
6. Cassandra;
7. Microsoft Access;
8. Redis;
9. SQLite.

In total, 7 out of 10 representatives of the rating are relational databases, as well as one instance of a document-oriented database (MongoDB), with distributed values ​​(Cassandra) and using the key-value approach (Redis). Thus, today the dominance of relational databases is undeniable, but what will happen tomorrow?

To answer this question, let's turn to the trends section on the same resource. If we take time stamps of more than 2 or 4 years, then the graph theory approach demonstrates the greatest growth. At the same time, over the past year, databases based on time data have shown the maximum growth in popularity. This is a relatively new approach, it is also considered NoSQL, the advantage is to create a structure based on dates or time ranges. At the moment, the most popular representative of the Time Series database is InfluxDB.

What databases do you use? And what is the most promising NoSQL database in your opinion?