Shulou(Shulou.com)05/31 Report--

This article mainly explains "how to use time series database". The content of the explanation in this article is simple and clear, and it is easy to learn and understand. let's study and learn "how to use time series database"!

A brief introduction to time Series Database-InfluxDB,TimescaleDB and QuestDB

Based on specific business requirements and use cases, a large number of new databases are derived from relational databases. Key values are stored in memory to graphic databases, from geospatial databases to time series databases. In cases where the general solution for using relational databases is not very effective, all of these different types of databases have specific uses.

Although there are many different types of databases, here we will study the time series database-the database needed to process time series data.

The data consisting of continuous measurements of something during a time interval is time series data.

With the modernization of financial transactions and the emergence of the Internet of things, the demand for time series database is obvious. The prices of stocks and cryptocurrencies change every second. In order to measure these changing data and analyze them, we need an effective method to store and retrieve the data.

With the unprecedented penetration of Internet of things devices in our lives, the data generated by Internet of things devices is increasing every day. Whether it's car diagnostics, temperature readings of houses, GPS locations of stray dogs, Internet of things devices are everywhere.

Internet of things devices can only do one thing, one thing. The information is captured by the sensor on the device and sent to the server for storage. Because the existing communication protocols are too complex for such lightweight high-frequency data and streaming data, MQTT is developed to solve the message delivery of the Internet of things.

There are two types of time series data-regular (usually based on measurements) and irregular (usually based on events).

But time series data is not limited to the Internet of things, it also permeates the entire Internet. Time series data can also be generated by capturing trends such as search engine queries, topic tags, and viral social media posts. It doesn't end there. In a software-driven world, documenting and reviewing security and compliance is critical. All these data can also be classified as time series data.

A time series database is specially designed to deal with problems caused by capturing, storing and analyzing time series data from one or more of these sources. So, for simplicity, let's define the time series data as:

High capacity (continuous data from measurements)

Natural chronological order (time is an essential dimension)

More valuable as an entire dataset (than a single record)

With this information, the time series database should be able to store large amounts of data and have the capabilities of large-scale record scanning, data analysis and data life cycle management. As mentioned earlier, although traditional transaction databases can be used to store, retrieve and process time series data, they can not make full use of available resources.

Specific problems require specific solutions.

Now, as companies realize this fact, they are starting to use specialized databases to solve specific problems. Now it's time for me to start what I'm talking about in this article. Among all other databases, the time series database has had a higher adoption rate over the past two years (data as of December 2020).

The 2.5-fold increase in time series database usage is mainly due to the convergence of cloud and data technologies, as well as the ability to capture data from earlier places, that is, car engines, refrigerators, location data for billions of devices, and so on. In addition to the new sources, the company realized that some of the older sources were not really suitable for transaction databases after all. All these promote the wide adoption of time series database.

Time series database

Considering the existence of a time series database, let's examine the different options that can be taken if we try a time series database. A complete list of time series databases can be found on the DB-engines website. I'm going to talk about three of them.

Time series database

As a PostgreSQL sale of a time series, it quickly caught your attention. By default, PostgreSQL is a compliment to everything. With the help of new architectures such as supertables and chunks the insertion performance of TimescaleDB has been improved more than 15 times and the query performance has also been significantly improved. Read more about it here.

Although major cloud providers do not have a fully integrated solution for TimescaleDB in the cloud, as most other time series databases do, TimescaleDB can run seamlessly on all of these databases. For example, if your infrastructure is in AWS and you do not want to run your TimescaleDB instance in Timescale Cloud, you can use the EC2 instance to install the official TimescaleDB AMI, or you can use AWS Elastic Kubernetes Services with the official helmet image.

InfluxDB

Unlike TimescaleDB, which is inspired by PostgreSQL (relational database), this database is a NoSQL time series database written from scratch. Although TimescaleDB has the advantage of standing on the shoulders of widely accepted and admired relational databases, InfluxDB has taken a different path. InfluxDB is a leader in time series database, but according to the research of TimescaleDB, it can not beat TimescaleDB in many aspects.

If you want to read interesting articles and want to install both databases on your system to learn for yourself, please go to this interesting comparison posted by Oleksander Bausk on his blog today.

Having said that, InfluxDB has many functions. In addition to the query languages InfluxQL and Flux, InfluxDB has developed a clean, lightweight, text-based protocol for writing points to a database. To its credit, other time series databases, such as QuestDB, have adopted this approach.

Like TimescaleDB, InfluxDB provides a cloud solution out of the box, but you can still decide to run InfluxDB on one of the cloud platforms. For example, if you run it on AWS, you will have native support for CloudWatch metrics, Grafana,RDS,Kinesis, and so on. All in all, a very good database. Because it is quite new, it is difficult to say that it will compete with temporal databases based on relational databases.

QuestDB

QuestDB is the latest member in the list of time series databases, and it is one of the latest batch of YCombinator. Some of the main differences in QuestDB are column storage, low memory footprint, the use of relational models in time series, and scalable schemaless extraction. Like most time series databases, QuestDB also uses official AMI and Kubernetes Helm Charts to provide cloud deployment options on AWS.

QuestDB also uses InfluxDB Line Protocol for reception without having to worry about changing the schema as the data structure changes. As a column database, QuestDB seamlessly handles the creation of new columns, so it supports schemaless extraction.

Although in the early days, QuestDB provided almost full support for ANSI SQL and some additions to the SQL dialect, it created many completely unique features that made it a viable alternative, possibly better than some other major databases on the market.

Thank you for your reading. the above is the content of "how to use time series database". After the study of this article, I believe you have a deeper understanding of how to use time series database. the specific use of the situation also needs to be verified by practice. Here is, the editor will push for you more related knowledge points of the article, welcome to follow!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.