tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Carlos Pereira <khep...@portugalmail.pt>
Subject RE: File writing performance
Date Wed, 22 Oct 2003 13:54:13 GMT
> The main problem with performance comes from your source code.  You are
> reading 1 byte at a time then writing 1 byte at a time.  Use a buffered
> input stream, and also use a buffered output stream
(Wade Chandler)

> Really efficient values are FS blocksize (linux defaults to 4096b).
(Francois JEANMOUGIN)

> (Have a look at the source of the DefaultServlet)
(Ralph Einfeldt)

 Ok, I'll use a 4096b long buffer. The code i posted was just a test; it worked, now i'll
go for performance tuning. My doubt was if it would be slower that the processing that the
default servlet does. After watching it's source i guess it will even be quicker, because
it doesn't need so many tests.

> There is another problem: Are you shure that flash is using 
> a simple download and won't use things from http 1.1 like 
> byte range requests (and won't do it in the next version).
(Ralph Einfeldt)
 I don't understand what you are talking about. SWF files work just like java applets; they
are downloaded, the flash player is started and runs the SWF file (much like the java plugin).
I really don't know anything about byte range requests. According to this behaviour, do you
think there might be any problem?

 I've searched the SWF file format specification, macromedia's developer center and Flash's
help files. They don't have any reference to that either.
 I tried an example and it worked fine, with no problems at all. But, i want to make sure
i don't need to worry about this...

Best regards,
Carlos Pereira

---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org


Mime
View raw message