SSDB with a single instance supporting hundreds of millions of requests per day

SSDB is a NoSQL storage server developed by C ++. It supports zset and map data structures and can replace Redis. It is especially suitable for storing collection data. after SSDB is developed and open-source, it has been tested in the production environment for three quarters and has been running stably. data in a list supporting tens of millions of users (such as the user's order history and user's friend column)

SSDB is a NoSQL storage server developed by C ++. It supports zset and map data structures and can replace Redis. It is especially suitable for storing collection data. after SSDB is developed and open-source, it has been tested in the production environment for three quarters and has been running stably. data in a list supporting tens of millions of users (such as the user's order history and user's friend column)

SSDB is a NoSQL storage server developed by C ++. It supports zset and map data structures and can replace Redis. It is especially suitable for storing collection data. after SSDB is developed and open-source, it has been tested in the production environment for three quarters and has been running stably.

SSDB processes hundreds of millions of read/write requests every day on an instance that supports the list data of tens of millions of users (such as the user's order history, user's friend list, and user's message list, the CPU usage remains at around 3%, and the memory usage is 1 GB. this data scale cannot be met by the original Redis, because Redis cannot store such a large amount of data, and the physical memory capacity limits the Redis capability. based on our experience, Redis is suitable for 10 Gb of data. When the data size expands, Redis is very difficult and can hardly be expanded. in this case, you must use SSDB instead.

SSDB has APIs that are highly coincident with Redis, and hash (map) can be traversed in segments. In comparison, Redis can only traverse all elements in hash at a time through hgetall, in a large hash, this operation is very inefficient.

If you want to list several items that must be abandoned in Redis and use SSDB, I believe these items are very attractive:

• The storage capacity of a single instance is equivalent to 100 Redis instances!
• The memory usage is only 1/1000 of that of Redis (max design capacity ).
• All data sets (including KV) can be traversed in segments (by PAGE.
• It is particularly suitable for storing list and other collection data.

SSDB is an open source project (https://github.com/ideawu/ssdb), you can get its source code for free, and you can start the server without programming and modifying the configuration file.

Related posts:

1. Key_range of SSDB and the path of future Clusters
2. PHP script for Redis Data Import
3. Use Cases of SSDB in log analysis of large data volumes
4. SSDB supports flushdb commands to clear Databases
5. SSDB adds the Compaction speed limit Function

The article you are reading is: SSDB that supports hundreds of millions of requests per day for a single instance

Linode VPS-web hosting in the united states | IT cattle blog website Aggregation