KeepAlive
| Description |
Enable HTTP Keep-Alive to serve multiple requests within a single TCP/IP connection |
Synopsis
|
KeepAlive [on | off] |
| Context |
Default Server, Virtual Host |
Example
|
KeepAlive on |
The KeepAlive directive enables the HTTP Keep-Alive facility where a single TCP/IP connection may be reused for multiple requests. They can significantly speed up overall performance particularly over longer connections where the TCP/IP connection setup time is long.
Keep-Alive is supported by AppWeb with both HTTP/1.0 and HTTP/1.1 protocols. With HTTP/1.1, Keep-Alive is the default. With HTTP/1.0, clients may optionally request Keep-Alive.
Connections can only be reused if the length of the content returned to the client is know ahead in advance to that a Content-Length header can be correctly formatted. This means that very large CGI, ESP or EGI output may cause a TCP/IP connection to be closed and the effectiveness of Keep-Alive to be limited.
KeepAliveTimeout
| Description |
Set the time period to hold a TCP/IP connection open for further requests when using Keep-Alive. |
Synopsis
|
KeepAliveTimeout seconds |
| Context |
Default Server, Virtual Host |
Example
|
KeepAliveTimeout 120 |
The KeepAliveTimeout directive defines the time period to keep connections open waiting for further requests from a given client. If further requests arrive, the timeout period is restarted.
Setting the timeout period too high may impact performance on high-traffic servers because of the large number of open TCP/IP connections.
MaxKeepAliveRequests
| Description |
Define the maximum number of requests for a connection. |
Synopsis
|
MaxKeepAliveRequests number |
| Context |
Default Server, Virtual Host |
Example
|
MaxKeepAliveRequests 50 |
The MaxKeepAliveRequests directive defines the maximum number of requests that may possibly be services using a single TCP/IP connect. Limiting the number to a reasonable maximum can help reduce the impact of some denial of service attacks.
SendBufferSize
| Description |
Set the TCP/IP buffer size. |
Synopsis
|
SendBufferSize limit |
| Context |
Default Server |
Example
|
SendBufferSize 8192 |
The SendBufferSize directive sets the TCP/IP buffer size. Increasing the SendBufferSize can improve throughput and lower CPU usage. You should only need to do this if you have very high data rates.
The
limit may be set to 0 to use the O/S default value.
SessionTimeout
| Description |
Set an inactivity time period before disposing of a session data store
|
Synopsis
|
SessionTimeout seconds
|
| Context |
Default Server, VirtualHost |
Example
|
SessionTimeout 1800
|
The SessionTimeout directive defines the number of seconds of inactivity to wait before disposing a session data store. Embedded Server Pages (ESP) will create session data stores to store user's state data. If the user abandons their session, the session will be cleaned up when this timeout expires. It is good practice for ESP pages to explicitly dispose of sessions when a user completes a transaction or logs off.
Timeout
| Description |
Set the time period to wait before cancelling the request |
Synopsis
|
Timeout seconds |
| Context |
Default Server, Virtual Host |
Example
|
Timeout 600 |
The Timeout directive defines the number of seconds of I/O inactivity to wait before cancelling a client request. This timeout is necessary to cleanup idle connections.
Some web user interfaces with monitoring applets open HTTP connections and keep them open to stream response data back to the applet. To prevent AppWeb from closing the connection, you may need to either increase the timeout value, or you may need to do some I/O to the applet to ensure it does not get closed.
Performance Directives