# Requirements for HTAP

+ Extensibility
  + Distributed Transactions
  + Distributed Storage
+ Support both OLTP and OLAP simultaneously
  + Support both row-based storage and column-based storage at the same time
  + OLTP and OLAP business are isolated.
+ Real-time
  + Real-time synchronization of row-store and column-store data

------------------------------------------------------------------------------
Based on the definition of HTAP and the problems it aims to solve, we propose the following requirements for a database supporting HTAP technology:

Regarding scalability, due to the large amount of data, distributed storage is a storage solution for OLAP business for some NoSQL databases. However, due to the need to support OLTP business at the same time, the requirements for real-time data writing and updating and distributed transactions have been raised.

At the same time, supporting both OLTP and OLAP business is necessary due to the different characteristics of the two businesses, but OLTP requires high concurrency and consistency, OLAP requires high-performance data scanning, and how to ensure that the two do not affect each other (business isolation) is also a problem to be solved.

Real-time data analysis has already become a common requirement, so can the row store service that undertakes real-time data writing transform and synchronize the data to the column store service in real time, which is also a problem that HTAP technology must solve.

In conclusion, a system that meets the HTAP scenario must be able to satisfy the above three points to be truly considered as a solution for the integration of OLTP and OLAP business.