Best databases of the year 2017

The database ecosystem, like any ecosystem, is subject to the iron rule of digital Darwinism, which posits continuous evolution and adaptation to the IT environment. Last time we looked at the best databases of 2016, and, although 12 months is not a long time in evolutionary terms, we’re back again to check out the hottest databases of 2017! That’s right, it’s the swimsuit issue! (Of databases)

If you’re in the market for a new database, the possibilities are basically the relational type on one hand, and on the other, the NoSQL variety. The earliest databases to emerge from the primordial digital ooze back in the 1970s dominated the IT world: they were large hungry beasts that devoured all the data they could sink their teeth into, until they began to die out and a new breed of database evolved, NoSQL databases, relatively recently in world history, in the last decade or so. Finally, like early hominids, these two distinct subtypes produced hybrid creatures, such as the SQL/NoSQL database, in-memory databases or DBaaS.

A selection of storage technology that is easy to use, secure and with practical tools as standard, plus a support community contributing to the lifecycle of the product, will all make a difference when it comes to managing your business and improving productivity, becoming a fundamental component of your business. Why leave the correct functioning of your database (and your network in general) to chance? Database monitoring is a priority in any IT installation.

If you were going to hire a new technician for your support team you would first check out their résumé and hold a preliminary round of interviews. Well, it’s no different when you’re looking for a hot new database for the office; you have to look at the pros and cons, and not only technical specs. You need to ask yourself: Where is the business at right now? Are we growing? Consolidating? Are we ready to scale up? Or to scale sideways? Will the database you choose adapt itself to the ecosystem in which it finds itself, or will it be a digital Dodo, unable to evolve quick enough to survive?

The essential questions are:

• How many clients do you need to service?
• What amount of data will you need to manage?
• Will you need to implement “batches” that access the database?
• What kind of response time do my clients expect?
• How will the database scale if my number of clients and transactions grows?
• How will you monitor your database to avoid down time?
• Do you need a relational database or NoSQL?
• Will the database misbehave if it crashes? How will it behave in case of a problem?

Follow the link to check out a comparative of NoSQL vs SQL databases.
And now, the moment you’ve all been waiting for, the loveliest models of 2017 are here, with their finest characteristics on display for all to see.

Hottest commercial databases

The market is presently dominated by DB2, SQL Server, Oracle and IBM. On Windows OS the SQL Server is the usual database of choice, while Oracle and DB2 are the apex predators of the Mainframe/Unix/Linux ecosystem.

Microsoft SQL Server

Developed by Microsoft, and exclusively compatible with Windows. Many are trained in its use, and it is easy to acquire. Since its integration with Microsoft Azure its flexibility and performance have improved, plus it now permits information from other servers to be administered, improving its usability.

Benchmark: http://www.microsoft.com/en-us/server-cloud/products/sql-server-benchmarks/industry.aspx

Oracle

Oracle can run on practically any system, and is served by many who are trained in its ways, having many adepts. Also of note are its many tools oriented toward monitoring and admin.

Oracle Benchmark: http://www.oracle.com/us/solutions/performance-scalability/index.html

IBM DB2

The second-most used database for Unix/Linux ecosystems, right behind the dominant Oracle, and the #1 choice for Mainframe. DB2 has its followers, trained in its arts, but there are fewer initiated in its ways than for Oracle. On the other hand, one who follows IBM D2 has no need to be able to operate in a Unix/Linux environment.

DB2 Benchmark: http://www-01.ibm.com/software/data/db2/performance.html

Teradata

Designed with Big Data in mind, its storage and data analysis capacity are gargantuan.

SAP Sybase

A survivor from an earlier age, when it reigned over its competition, it remains a solid bet in terms of scalability and performance.

Informix

Another victim of the Dot-com extinction event of the 1990s. After an initial positive run it got devoured by a large predator (IBM) following a series of bad management decisions. Like many technological evolutionary dead ends (see also Ascential) it has become extinct, although its vestigial digital DNA can still be found in some IBM tools and applications.

Hottest Open Source relational databases

In this category there are three outstanding beauties: MySQL, MaríaDB and PostgreSQL. While each one is attractive in its own right, they share some characteristics: a powerful support community, open code that users can modify as needed, and last but not least, that they are free and available.

NoSQL databases

NoSQL databases evolved in response to external conditions and, although they belong to the same family group, they have many unique characteristics, developed in isolation from others of their species. The family group has split into sub-categories that each model data according to their particular evolutionary processes. We can classify them in four groups:

NoSQL Oriented to key-value

The simplest data model, ideal for when you access your data by key. The difference being that in this case data can be stored without defining any specific schema. They are very efficient for reading and writing, and are designed to scale massively thereby achieving extremely fast response times. Data is usually stored in complex structures such as the amusingly named BLOB. Some examples of this kind of database would be the following:
– Redis: Open Source and free software.
– Riak: A dedicated key-value database, outstanding document storage and search functionality
– Oracle NoSQL
– Microsoft Azure Table Storage

Document oriented NoSQL

This system uses various formats (JSON, XML) and features the capacity to change schema without stopping the database. Developers can upload indexed documents and get access through the database storage engine. Their flexibility makes them one of the most versatile tools, Mongo DB and Couchbase Server being two noteworthy examples.

Mongo DB

One of the most flourishing databases of the present epoch, capable of working with both structured and unstructured data, and with excellent scaling and performance capabilities. It has a highly skilled and powerful coven of followers who can introduce the uninitiated into its ways. Will work with key-value pairs, allowing access to different parts of the stored data.

Mongo DB does not support operational atomicity but does guarantee eventual consistency. Changes are reproduced throughout all the nodes although it cannot be guaranteed that all nodes will receive the data at the same time.

Couchbase Server

An open-source database licensed under Apache. Despite its name it is unable to guarantee 100% data integrity. On the plus side it features an intuitive administration console, via which it is possible to easily access ridiculous amounts of data.

Mark Logic Server

This small, furry, seemingly insignificant database has triumphed thanks to its data integrity and XML, JSON and RDF compatibility.

Supported systems: Windows, Solaris, Red Hat, Suse, CentOS, Amazon Linux and Mac OS.

Elastic Search

Worthy of mention: RavenDB, Apache Jena and Pivotal GemFire.

Column oriented databases; No SQL

This species represents data values as “columns” that allow the user to map keys to values and group them in structures. They are used in environments where you need to access various columns with many rows. They are most useful for processing and events analysis, content management and data analysis.

Apache Cassandra

Created by Facebook and now freely distributed. Recommended for databases handling unimaginable quantities of data. There is also an Enterprise version called Datastax Enterprise.

Supported formats: ASCII, bigint, BLOB, Boolean, counter, decimal, double, float, int, text, timestamp, UUID, VARCHAR and varint.

Benchmark: http://vldb.org/pvldb/vol5/p1724_tilmannrabl_vldb2012.pdf

Apache Hbase

Designed to support multiple read and write access to huge amounts of data in real time. Hadoop mail and file system are two points in favor.

NoSQL graph databases

These models are focused on properties and the relations among them, making use of graph theory to connect databases. Each element is joined to its adjacent element. Recommended if your data is highly inter-relational, as in social media networks, fraud detection, real time updates, etc. The logic is, a column for each structure, and for each relation two columns.

Neo4j

Supports data integration, high availability and clustered scaling, with furthermore a good administration panel.

Infinite graph

A license-only database that supports Mac OS, Linux and Windows.
Benchmark: On demand from Objectivity.

Hybrid models

More companies are offering hybrid solutions that use various database engines that admit various NoSQL setups plus relational engines.

To name a few examples, these include, CortexDB, Foundation DB and Orient DB, which all offer different NoSQL models.
IBM has extended its DB2 databases to include the possibility to use BLU Acceleration with NoSQL database, allowing data to be saved in XML, JSON and graph mode.

Databases as a Service

Database as a Service’s offer is differentiated from the rest of the formats in this article as it is Cloud-based. The user simply inputs the data to be saved and the service provider does the rest. The convenience of this system is sure to increase its popularity in the future.

Amazon SimpleDB

A database that offers a simple web service interface to store and create data sets. If you want to create access to simple databases Amazon SimpleDB is an option to keep in mind.

Data is saved as text and generates structures made up of pairs of parameter values. Data is indexed automatically, making searches very convenient.

Benchmark not available.

List of Pandora FMS database monitoring modules:

This article is written as a quick introduction that hopefully points out the need for a prior study of your own company or organization’s requirements. It’s not necessary to invest in the most gargantuan, fastest, vertically scaling solution if your business does not demand those characteristics. Talk to your tech people and work out which is the best solution for the short to medium term.

Have we forgotten your favorite database? Let us know if your perfect solution didn’t make the list and we’ll include it in our end-of-year round up when we distribute our in-house awards for best tech products of the year, the Doras.

If you’d like to know more about Pandora FMS and its database monitoring potential, or about any of the other many, many things our flexible monitoring solution can offer, just follow the link.

Pandora FMS team

El equipo de redacción de Pandora FMS está formado por un conjunto de escritores y profesionales de las TI con una cosa en común: su pasión por la monitorización de sistemas informáticos.

Pandora FMS’s editorial team is made up of a group of writers and IT professionals with one thing in common: their passion for computer system monitoring.