How do I connect to CTSDB?
Is CTSDB easy to use?
Yes. In the console, you can view the instance details, perform management and control operations such as initializing and renaming instances, and use the monitoring feature to view the instance health status in real time. You can also use RESTful APIs to perform almost all types of data operations, and such APIs are fully compatible with Elasticsearch protocols.
It implements an inverted index algorithm to accelerate queries in any dimension.
What are CTSDB's advantages over relational databases?
In scenarios with massive amounts of time series data, a relational database will have the following problems:
High storage costs: Its performance in time series data compression is poor, and it needs to use large amounts of machine resources.
High maintenance costs: It uses a standalone system, where databases and tables need to be partitioned manually at the upper layer, incurring high maintenance costs.
Low write throughput: Its write throughput based on a single server is low, which can hardly withstand the pressure of writing tens of millions of time series data points.
Poor query performance: It is more suitable for transaction processing but has poor performance in aggregate analysis of massive amounts of data.
In contrast, CTSDB has the following advantages:
Low storage costs: Considering that the data is usually incremental in time, has repeated dimensions, and has smooth metric value changes, it uses appropriate encoding and compression algorithms to increase the data compression ratio and aggregates historical data through data rollup to save storage space.
Highly concurrent writes: CTSDB adopts the policy where data is written into the memory first and then periodically dumped as immutable files for storage. It can also write data in bulk to reduce network overheads.
Low query delay and high query concurrency: It optimizes common query modes, uses technologies such as indexing to reduce query delay, and leverages technologies such as caching and routing to improve query concurrency.
How do I connect to CTSDB on a CVM instance over the classic network?
Currently, CTSDB can be accessed only in VPCs. If your CVM instance is on the classic network, you can interconnect the classic network with a VPC through Classiclink to connect to CTSDB.