Pika is a Redis storage system that is jointly developed by the DBA and infrastructure groups, fully supports the Redis protocol, and allows the user to migrate services to Pika without modifying any code. DBA Maintenance Pika with Redis experience does not require learning costs.
Pika main solution is the user using Redis memory size of more than 50G, 80G and so on, will encounter a long startup recovery time, a major from the cost of large, expensive hardware, the buffer is easy to write full and so on. Pika is a solution to these scenarios.
Pika is now open source, GitHub address:
Https://github.com/Qihoo360/pika
Focus
Pika's single-threaded performance is certainly not as redis,pika as the multi-threaded structure, so the performance of some data structures can be better than Redis in the case of more threads.
Pika is certainly not completely superior to Redis schemes, but is better suited under certain scenarios. So at present the company internal Redis,pika is the common existence plan. The DBA chooses the right scenario based on the business scenario.
360 Open Source Class Redis storage System: Pika