Architecture scheme: the technical architecture scheme of a logistics screening single system with an average daily traffic of 0.5 million? For a logistics screening system with an average daily traffic of about 0.5 million, the system must respond within 3 seconds and have requirements on system reliability, stability, and scalability. what technical architecture is recommended? Architecture scheme: the technical architecture scheme of a logistics screening single system with an average daily traffic of 0.5 million?
For a logistics screening system with an average daily traffic of about 0.5 million, the system must respond within 3 seconds and have requirements on system reliability, stability, and scalability. what technical architecture is recommended? System architecture:
------ Solution --------------------
There are more than one nginx and server.
The key is the bottleneck. if the server doesn't work, the machine will be added. if the database doesn't work, for example, using a memory table, the memory will be added. Otherwise, the database and table will be split. The key is to view the business requirements and identify the bottleneck.
------ Solution --------------------
First, test the system bottleneck and then go to the next step.
------ Solution --------------------
The server uses lnmp, and the database uses table sharding by month. if the money is not a problem, perform a Database read/write splitting.
------ Solution --------------------
Reference:
I checked some information on the internet. my initial idea is to use LNAMP and static resources to enable nginx to provide load balancing at the same time. apache (use two servers first) is used as a cluster to handle php, the database mysql (2) is used as a cluster, and the database cache uses memcached. then, the system bottleneck is tested and adjusted.
Quote: reference:
The server uses lnmp, and the database uses table sharding by month. if the money is not a problem, perform a Database read/write splitting.
The database mentioned on the third floor adopts monthly table sharding or read/write splitting, which is not very clear. can you elaborate.
1. Can a logistics screening list be understood as logistics order data? If yes, you can split the data into sub-tables.
2. there is no need for static servers, because you are querying data, not images.
3. there is no problem with one server in your system, and the two servers are more stable.
4. if you use two servers, I suggest using the database for read/write splitting (or master/slave). The web program can use two services for load balancing.
My opinion is whether daily access is PV or UV, even if it is done by two UV servers. This is enough. The data size is too small.
------ Solution --------------------
It is better if you use Linux + php + nginx + redis. you can use mysql as the backup database. the memory size will be higher.
