# Architectures Comparison

__*TiDB architecture advantages*__

+ Compared with traditional non-distributed database architectures, TiDB has the following characteristics:
  + Both support ACID and strong transaction consistency
  + Distributed architecture, decoupling of components, good scalability, supporting elastic expansion and shrinking of the cluster 
  + High available by default, fully automatic fail-over, and the database can automatically fail over in the event of a few failed replicas, transparent to the business
  + Horizontal scaling, without the need to change business logic or add sharding keys etc
  + The strength is not in the responsiveness of lightweight simple SQL, but in the throughput of a large number of high concurrency SQL 
-------------------------------------------------------------
Under the premise of ACID compliance, the main difference between a traditional database and a TiDB cluster is that the former is designed to run on a single-instance, and the most suitable scale method is vertical, but horizontal scaling is very difficult. Once a multi nodes cluster deployment is involved, the communication cost of the nodes increases dramatically and the architecture tends to be complex; the latter is a distributed multi nodes design, with excellent horizontal scalability. This is one of the reasons why it's called cloud native.

In contrast, the characteristics of the two are clear. In the case of small data volume and very simple operation, the single-instance architecture is sufficient, and there is no additional network communication overhead, the effect is ideal; but in the case of massive data, update operation throughput is huge, the advantage of TiDB is obvious, and node can be easily added to cope with the increasing data volume close to the trend of linear growth. and performance challenges.