Four-tier architecture for distributed database servers

Four-tier architecture for distributed database servers:

Access layer: Receives access information and distributes it intelligently to the staging server, accepting data results and returning to the client.

Transit layer: Receive the access server sent data access instructions, from the total storage server to find the data distribution of the storage server, send instructions.

Header layer: Stores the header information of the data to determine the location of the storage server.

Processing layer: A distributed data storage server that receives instructions and executes, and then returns data to the Access server.

function Distribution:

The access server only does four things: receives the client's access data, receives the load status information of the transit server, and assigns the data to the lowest load
Relay server, and returns the client after receiving the result.

The relay server only does four things: it is responsible for receiving access data, accessing the Header table Server query location, receiving the results, and then passing the instructions of the Operation data
to the processing server.

Header server only do four things: Store total data header, receive query data, find the server location of data, return location information to transit service
Manager

The processing server only does four things: store data, receive instructions, execute instructions, and then return the results to the access server.

Technical Brief:

"Transitive" and "reactive" are combined, the response as the basis, and passed as the judgment result. For example: The access server receives access to the data,
The server listens for events and responds and returns the load status, the server that accesses the server to determine the lowest load passes its data; the header server receives
A query request that governs the processing of the server response data and returns whether the table header server is passed to the staging server based on the presence of the data
Information, the relay server determines whether to continue querying other header servers based on the response, and the process can be parallel until there are definite results
The query is aborted.

Architecture Summary:

As long as there is demand, theoretically can infinitely increase the level of the server to deal with.

Four-tier architecture for distributed database servers