High-concurrency and high-traffic website architecture (Internet products)

 

1. Database Layer

ORACLE: Table Partitioning, historical data migration, multi-field merging (defining compression rules), rational connection pool, Id needs to be centrally allocated using common algorithms, Distributed Database Cluster load balancing
MySQL: historical data migration, multi-field merging (defining compression rules), rational connection pool, Id needs to be centrally allocated using a common algorithm, Distributed Database Cluster load balancing

Database Cluster:

Oracle's real application cluster (RAC)
Microsoft SQL Cluster Server (MSC)
IBM's DB2 UDB high availability cluster (UDB)
Sybase ASE high availability cluster (ASE)
MySQL high availability cluster (MySQL CS)

2. HTML static + Ajax

Freemarker: Static HTML in large regions

Ajax: the data that needs to be changed in a timely manner. The data can be pushed by messages or obtained through Ajax.

 

3. Server Load balancer and cache Cluster

Apache2 + Tomcat/apache2 + JBoss: Load Balancing/acceleration

-Djava. NiO. channels. SPI. selectorprovider = sun. NiO. Ch. epollselectorprovider (Tomcat supports epoll)

Memcached (it is best to disable Linux swap): cache Cluster

 

Lighttpd: distributed storage of small files

 

GFS: The Distributed File storage tool of RedHat, which has the advantage of transparent applications and high cost.

 

Search: Search for the required data and create an index file based on the keyword

 

4. Image Report Server Separation
Image report service and page service separation, image Report Server Load Balancing