What is the difference between client-server and distributed architecture

[avit]

Hi,

I understand that the internal architecture of each component, the domain model, and so on .... but I want to know, is there any difference between distributed architecture and Client / Server, does not represent the same thing? Someone say that past is Client/Server application , and now is distributed application, but i don't understand these terms.

[JIII]

Under the scheme distributed architecture, when a terminal requests to the server, the requested information must travel across the network to be processed locally by the workstation, that is, if you have a 10MB database and run a search (SEEK or LOCATE) on it which does not report any data found, all data, the information contained in 10MBytes DBF file will travel the network cable to the workstation that requested the search. This transfer Each terminal generates massive collisions and low network performance, the less you have your network speed (10 mbps or in the case of a VPN, for example), and there is more distance between terminal and server, the application will take longer to access to data, with the usual loss of time and slowness of the programs.

Under this model of work, when a terminal requests to the server, the requested information is processed by the data server (which is a program that runs on the file server and monitors all the time requests for data files), This data server locates the requested information and only the response information is traveling over the network, ie the example above, where we have a DBF of 10MB and runs a search on it in which no data only travel a few bytes to the terminal by the network cable, indicating the results of the search, this has the effect of reducing collisions by 90% or more, depending on the application, thereby increasing the access speed data.

[Theodot]

In the case of distributed architecture, the server opens the data files for each terminal 1 time you apply, this means that if the application opens 10 data files and if we had 10 terminals accessing the same data, the server would open 100 files. In case of failure in any of the terminals receive the famous error of corrupt indices (corruption detected) that as much notoriety has given DBF files but that is not a problem of DBF file structure, but the same distributed architecture on which are built local area networks, any computer system, made in any programming language with any file format that does not use a client / server technology is subject to such problems, regardless the type of data file handling or in which language has been developed.

In the client / server, data server (software running on the file server) is responsible for controlling the opening of the databases, which are opened only once, regardless of the number of workstations that are accessing the files. In case of failure in any of the terminals, the data server controls that do not close the files that are being used by other users, and in case of widespread failure in all the terminals is the same server that closes all database appropriately and thus greatly reduces the possibility of corruption of data or index files always leaving data in its last known state.

[Una]

Under a scheme client / server, fully exploits the responsiveness of modern processors because it is the same server that performs operations to access data, thus the server response speed is proportional to its capacity process.

Another problem that brings distributed architecture is the waste of server processing power, since in this model, the network server only acts as a shared hard disk and not making the stunning information processing capabilities at their disposal the next-generation processors such as Xeon, Pentium IV, Athlon, etc.. as a result, the server response rate is not proportional to its processing capacity, which means that no matter how powerful is our server, problems of slow access to data will remain

[shrnmlss6]

Client Server - As the name says you have a client and a server, client will ask for some service, which the server will deliver.

Distributed Architecture - This is when we have more than 1 machine involved in system. Say for a ecommerce project you may have a app server which will server UI, but you could have some other server to process orders. You could have more than 1 database. Client server is distributed to an extent. But usually it is tied to 2 or 3 tier. Distributed could be n-tier.

Tier - Physical seperation DB_Server01, APP_Server01 and WEB_SERVER01.

Layer - Logical seperation like Business layer, common layer, data acess, web components, windows interfaces...

[D_chapple]

Thank you so much shrnmlss6 for such a easy explanation of difference between client-server and distributed architecture. I hope that this will be helpful to all others who are searching for the same.