Server Design Alternatives

Ø The Client

Ø Iterative Server

Ø Forking on Demand

Ø Pre-Forking

Ø Threading on Demand

Ø Pre-Threading

Ø The Client

v client.c

Example Use:

% client localhost 10123 5 500 4000

Means:

· Connect to the server running at localhost port 10123.

· Forks 5 children.

· Each child makes 500 TCP connections.

· Each connection sends the string: "4000\n" and gets 4000 bytes from server.

Ø Iterative Server:

v serv00.c (it uses web_child.c & pr_cpu_time.c)

To test this server, we run only one client as:

% cd /home/cs779/stevens3rd.book/unpv13e/server

% serv00 10123
% client localhost 10123 1 5000 4000

When client is done, interrupt the server:

^C

and it will print

user time = 0.134386, sys time = 1.58777

Ø Forking on demand:

v serv01.c (it uses sig_chld_waitpid.c )

% serv01 10123
% client localhost 10123 5 500 4000

When client is done, interrupt the server

^C

Ø Pre-Forking:

v serv02.c (it uses child02.c )

% serv02 10123 3

/* 3 is the number of pre-forks /*
% client localhost 10123 5 500 4000

When client is done, interrupt the server

^C

Ø Threading on demand

v serv06.c

% serv06 10123
% client localhost 10123 5 500 4000

When client is done, interrupt the server

^C

Ø Pre-Threading:

v serv07.c (it uses pthread07.c  )

% serv07 10123 3

/* 3 is the number of pre-threads /*
% client localhost 10123 5 500 4000

When client is done, interrupt the server

^C